Re: GNUstep paste board (Was: really attracting developers)
Hi Matt,
it took some time, but I finally did get around to commit part of your
changes to SVN. I changed the code slightly and there still is a lot to
do, but it is a start for DnD interaction with non-GNUstep applications.
Cheers,
Fred
Matt Rice wrote:
--- Fred Kiefer [EMAIL PROTECTED] wrote:
DND will be another big step. Here a whole rewrite
is needed. When we
tried to integrate GNUstep DND with X we did get it
completely wrong the
first time.
a while back i tried getting gnustep to work as a x
'dragging source' as we call it.. it still uses the
outdated xdnd code but attached is what i ended up
with and a dragging destination which also uses the
outdated xdnd code from gnustep..
the whole thing inside the case SelectionRequest:
needs to be rewritten doesnt do any of the conversion
it should (file uris should have file:// etc and other
conversion methods for other stuff...)
i recall there being alot more to the patch than this
but this is all there seemed to be... so hopefully its
all here...
i also remember there being times when the code just
magically didnt work..
hope it helps, dont forsee finishing it anytime soon
if someone wants to pick it up...
__
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
Index: Source/x11/XGServerEvent.m
===
--- Source/x11/XGServerEvent.m(revision 22775)
+++ Source/x11/XGServerEvent.m(working copy)
@@ -30,6 +30,7 @@
#include AppKit/NSGraphics.h
#include AppKit/NSMenu.h
#include AppKit/NSWindow.h
+#include AppKit/NSPasteboard.h
#include Foundation/NSException.h
#include Foundation/NSArray.h
#include Foundation/NSDictionary.h
@@ -1294,10 +1295,34 @@
break;
case SelectionRequest:
- NSDebugLLog(@NSEvent, @%d SelectionRequest\n,
- xEvent.xselectionrequest.requestor);
+ {
+ DndClass dnd = xdnd();
+ /* ugly hack alert delete all this stuff please*/
+ NSPasteboard *pb = [NSPasteboard pasteboardWithName:NSDragPboard];
+ NSArray *types = [pb types];
+ NSLog(@%@, [pb types]);
+ NSData *data;
+ if ([types count])
+ {
+ id uh = [pb propertyListForType:[types objectAtIndex:0]];
+ if ([uh isKindOfClass:[NSArray class]])
+ {
+ if ([uh count])
+ {
+ uh = [uh objectAtIndex:0];
+
+ }
+ NSLog(@%@, uh);
+ data = [uh dataUsingEncoding:NSASCIIStringEncoding
allowLossyConversion:YES];
+ }
+ xdnd_selection_send(dnd, xEvent.xselectionrequest, (unsigned
char *)[data bytes], [data length]);
+ }
+
+ NSDebugLLog(@NSEvent, @%d SelectionRequest\n,
+ xEvent.xselectionrequest.requestor);
+ }
break;
-
+
// We shouldn't get here unless we forgot to trap an event above
default:
#ifdef XSHM
Index: Source/x11/XGDragView.m
===
--- Source/x11/XGDragView.m (revision 22775)
+++ Source/x11/XGDragView.m (working copy)
@@ -276,6 +276,7 @@
break;
case GSAppKitDraggingEnter:
+ xdnd_set_selection_owner(dnd, dragWindev-ident, typelist[0]);
xdnd_send_enter(dnd, dWindowNumber, dragWindev-ident, typelist);
xdnd_send_position(dnd, dWindowNumber, dragWindev-ident,
GSActionForDragOperation (dragMask operationMask),
___
Discuss-gnustep mailing list
Discuss-gnustep@gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
Nikolaus Waxweiler [EMAIL PROTECTED] [...] And why should a wiki not be appropriate for an app db? I'm sure I wrote suboptimal, not inappropriate. Applications have lots of metadata about them - some of it you can bodge into a wiki with Category tags and so on, but not much. Consider how much easier it is to find an application on Freshmeat than a whole-web search engine. What does the release or popularity of a project have to do with using a= wiki as the homepage? Volume of material to store, for one. The Amarok wiki looks fine to me. What exactly is = your problem with it? I don't find white on white to be a nice colour combination. Maybe your eyes work normally and you have your browser default to black-on-white? [...] Which browser are you using? What keeps you from accepting cookies just = for the wiki if you know it requires them? It depends which device I'm using, but the cookie requirements of the mediawiki seem not documented anywhere on it. Yes, I could step through and yes/no each cookie, but that's a PITA. Why don't you fix the wiki? I have never heard of P3P so I neither know nor care, but you can easily= copy the wiki text to some file and edit it offline :) Same way you'd ed= it anything else offline. This is part of the problem - WikiFarmers seem to be ignorant of web technology like P3P or WCAG and don't care. Most seem happy with the idea of survival of the fittest and see obstructing others are part of showing their fitness. Surprisingly given that, they also seem to be ignorant of history and unwilling to research, such as why the hostname was mediawiki.gnustep.org for a while. I edit stuff offline by synchronising online, editing offline, then synchronising again when I get back online. Copy-pasting between text files and browser is very primitive - it would nuke other later changes on upload, wouldn't it? That was a reference to how a wiki (here: Mediawiki) facilitates pointers to [[other article]]s on the same wiki. Well, I tried. That seems stupid of mediawiki to pick the []s traditionally used to mark editorial changes of quotes. Wiki uses CamelCase for links. Really, posting guides to gnustep.org (except as archive copies) is a poor substitute for getting articles about gnustep out there [...] What do you mean by archive copies? Stuff you can easily print out? I mean copies kept on gnustep.org in case the original goes off-line. I already stated that I have no problem with that per se -- it's just that= with writing a texinfo article, the author has to update it himself and = whatnot. If he loses interest in that or just vanishes or something, = someone else has to take over and blablabla. I think it's good that authors can use texinfo *if they want to*. If any author vanishes, there's going to be some pain to take it over, and a markup conversion is just another part of that. That said, texinfo has been around a lot longer than mediawiki's not-quite-wiki markup and is more consistent across platforms. No-one has to use texinfo for things on the web if they don't want to, though. One big plus of document-style markup is that we can format them into whatever is the current 'house style' without needing to enforce a tight style guide like discussed in http://lists.gnu.org/archive/html/discuss-gnustep/2006-09/msg00189.html Single-site wiki markup rules seem far worse than one author using texinfo. It's just easier to update information about moving targets when having them all in a central place= where more than one person can edit them. As noted before, it's harder to publicise. Also, the whole web already uses Wikipedia, so there's nothing wrong abo= ut that :O. Wikipedia is an interesting experiment, but terrible as an information source and people who rely on it make some hilarious mistakes. For one example, Wikipedia use resulted in an insult replacing a language name on some Elephants Dream DVD menus. http://slashdot.org/articles/06/07/16/1345246.shtml Then there are other problems like biggest-network-pipe-wins, the right-of-centre NPOV and attribution stripping. Hope that explains, -- MJR/slef ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
Applications have lots of metadata about them - some of it you can bodge into a wiki with Category tags and so on, but not much. Consider how much easier it is to find an application on Freshmeat than a whole-web search engine. Mh. Tags like [Translations] and [Environment] can be perfectly emulated with categories and then listed on a special page. Statistics not so much, but they're not essential anyway. But yes, this might become unwieldy when the number of GNUstep apps hit 1000 or so, I give in :O. The current appdb runs on WebObjects, good, but needs some updating of both functionality and look. So much to do, so little time... What does the release or popularity of a project have to do with using a wiki as the homepage? Volume of material to store, for one. Well... Wikipedia!!!1 ...?! I don't find white on white to be a nice colour combination. Maybe your eyes work normally and you have your browser default to black-on-white? So you're looking for a skin where colors and font-sizes are only set when necessary and you get to use your browser defaults? Try the Simple skin: http://wiki.gnustep.org/index.php?title=Main_Pageuseskin=simple. Set it in Preferences - Skins. Also, if you use Firefox or Opera, you can set specific user stylesheets for certain (or all) sites manually. I forgot how, though. It depends which device I'm using, but the cookie requirements of the mediawiki seem not documented anywhere on it. Yes, I could step through and yes/no each cookie, but that's a PITA. Why don't you fix the wiki? Hum. According to Google, Mediawiki supports logging in for a session, so I guess there's a way to avoid cookies. What happens when you don't check remember me? Try activating accepting of cookies, logging into the wiki and setting remember me, then deactivating accepting of cookies. You should stay logged in, but this is just a wild guess. Also, I'm no PHP dev :O. This is part of the problem - WikiFarmers seem to be ignorant of web technology like P3P or WCAG and don't care. Most seem happy with the idea of survival of the fittest and see obstructing others are part of showing their fitness. Surprisingly given that, they also seem to be ignorant of history and unwilling to research, such as why the hostname was mediawiki.gnustep.org for a while. Now you're sounding bitter -- they probably just didn't *know* some people had difficulties with their site, so you need to tell them explicitly. Don't just hint at non-WCAG-compliance :) And really, the hostname thing was a cosmetic issue. I edit stuff offline by synchronising online, editing offline, then synchronising again when I get back online. Copy-pasting between text files and browser is very primitive - it would nuke other later changes on upload, wouldn't it? Yes, but what other way is there short of having access to SVN and therefore being able to easily merge changes? It's the same with a static website or a content management system without such capabilities. Someone changes it while you're editing away and you'll have to use diff and/or patch. The chances of colliding might be less if you edit just a section, not the whole page. I mean copies kept on gnustep.org in case the original goes off-line. Wiki pages are kept on-site and are easier to publish/pass around because you know there aren't going to be outdated mirrors. That said, texinfo has been around a lot longer than mediawiki's not-quite-wiki markup and is more consistent across platforms. So is HTML generated by a wiki :O. Plus, you can adjust font sizes and whatnot. One big plus of document-style markup is that we can format them into whatever is the current 'house style' without needing to enforce a tight style guide like discussed in http://lists.gnu.org/archive/html/discuss-gnustep/2006-09/msg00189.html Single-site wiki markup rules seem far worse than one author using texinfo. Point 1, a hierarchy, relates to the structure (- navigation) of the wiki, not single articles. Point 2 really applies only to the application pages, because he didn't know that I originally intended to import the software infobox from Wikipedia (see the box to the right on http://en.wikipedia.org/wiki/Gnustep) which is a template that takes arguments. You don't have to care about layout when writing normal articles and whatnot. You structure the text, Mediawiki does the formatting for you. Wikipedia is an interesting experiment, but terrible as an information source and people who rely on it make some hilarious mistakes. Alright, lets stay with manual registration ;) ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
Nikolaus Waxweiler [EMAIL PROTECTED] What do you have in mind for which the wiki is suboptimal? For example, the applications database, the automated documentation and pretty much anything that's more structured than a text file. Wikis can be bodged into doing those jobs, but that's not optimal. Projects like Etoile and Beep Media Player use a wiki as their homepage and http://amarok.kde.org is mainly a front-end to it. Etoile has not released yet, I don't know Beep and http://amarok.kde.org/ looks confusing and hard to read here. And as I already said earlier, the spamming can be limited. Yes, there's the issue of defacement, but it won't happen that often if the software is kept up to date. 'spamming can be limited' = can't be avoided? 'if the software is kept up to date' - but can we keep the software up to date? Also, the wiki activity probably amounts to more than what the current homepage has seen recently. Agreed, but I think that's more down to misdirection by posters to this list suggesting that the wiki will be the main site; and that the limited access for most people to the non-wiki-website is almost as frustrating as access to the wiki site is for me. (The wiki is broken as it does not allow me to log in, seems to require cookies but doesn't document that or support p3p, doesn't allow offline working, and so on...) IMO, the biggest problem of the wiki is bitrot - if no-one has time soon to do the big updates of www that it needs, it sits there. But if no-one has time to do the big update of mediawiki that it needs, someone breaks in and announces the closure of the project :-/ [...] [...] If you can get it published in the community press, then that also helps to introduce new people to GNUstep. Please email link suggestions to [EMAIL PROTECTED] [...] New people are turned off by old and/or incomplete information. Besides, texinfo isn't the best format for linking to relevant information somewhere else, something a wiki does [[better]]. What word have you redacted with [[better]]? Are you quoting something? Really, posting guides to gnustep.org (except as archive copies) is a poor substitute for getting articles about gnustep out there in the wider world. I'm puzzled why you think it should force the whole web site into using a wiki. Regards, -- MJR ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
For example, the applications database, the automated documentation and pretty much anything that's more structured than a text file. Wikis can be bodged into doing those jobs, but that's not optimal. I already said that the automatically generated doc isn't going in the wiki. And why should a wiki not be appropriate for an app db? Etoile has not released yet, I don't know Beep and http://amarok.kde.org/ looks confusing and hard to read here. What does the release or popularity of a project have to do with using a wiki as the homepage? The Amarok wiki looks fine to me. What exactly is your problem with it? 'spamming can be limited' = can't be avoided? If the current system of manual activation of accounts is replaced with an automatic method, yes. Well, even then there's a small chance, but you can always revert changes. 'if the software is kept up to date' - but can we keep the software up to date? I'd like to think so. Updating is probably done by the hoster of gnustep.org afaik, you just have to tell them. Besides, nothing gets lost in a wiki :) Well, unless the SQLd blows up or someone deletes everything with some SQL injection. The hoster might have backups, though. Agreed, but I think that's more down to misdirection by posters to this list suggesting that the wiki will be the main site; If it doesn't become the homepage, at least have it become the foundation, no? I don't care either way as long as gnustep.org becomes a better place. and that the limited access for most people to the non-wiki-website is almost as frustrating as access to the wiki site is for me. (The wiki is broken as it does not allow me to log in, seems to require cookies but doesn't document that or support p3p, doesn't allow offline working, and so on...) Which browser are you using? What keeps you from accepting cookies just for the wiki if you know it requires them? I have never heard of P3P so I neither know nor care, but you can easily copy the wiki text to some file and edit it offline :) Same way you'd edit anything else offline. IMO, the biggest problem of the wiki is bitrot - if no-one has time soon to do the big updates of www that it needs, it sits there. But if no-one has time to do the big update of mediawiki that it needs, someone breaks in and announces the closure of the project :-/ So it's not bitrot but vandalism :P. Not that grave -- just revert it.. What word have you redacted with [[better]]? Are you quoting something? That was a reference to how a wiki (here: Mediawiki) facilitates pointers to [[other article]]s on the same wiki. Well, I tried. Really, posting guides to gnustep.org (except as archive copies) is a poor substitute for getting articles about gnustep out there in the wider world. I'm puzzled why you think it should force the whole web site into using a wiki. What do you mean by archive copies? Stuff you can easily print out? I already stated that I have no problem with that per se -- it's just that with writing a texinfo article, the author has to update it himself and whatnot. If he loses interest in that or just vanishes or something, someone else has to take over and blablabla. It's just easier to update information about moving targets when having them all in a central place where more than one person can edit them. Also, the whole web already uses Wikipedia, so there's nothing wrong about that :O. I'm not convinced. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
Nikolaus Waxweiler [EMAIL PROTECTED] [...] It has a nicely visible link to developer guides. The thing is, I need help linking all the guides/tools listed on http://www.gnustep.org/developers/documentation.html and http://www.gnustep.org/experience/DeveloperTools.html. The wiki should eventually replace the current main page (www.gnustep.org) and *porting* No, it shouldn't. There are some things that a wiki is suboptimal for, the wiki has only recently seen sustained activity and the wiki was used for the most recent defacement and spammings. all articles to the wiki would help this. BTW, new guides should be started on the wiki, not in some texinfo file :O autogsdoc and texinfo remain the recommended ways to create manuals, but html (however created) would also be fine for linking from www.gnustep.org. If you can get it published in the community press, then that also helps to introduce new people to GNUstep. Please email link suggestions to [EMAIL PROTECTED] -- MJR/slef ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
No, it shouldn't. There are some things that a wiki is suboptimal for, the wiki has only recently seen sustained activity and the wiki was used for the most recent defacement and spammings. What do you have in mind for which the wiki is suboptimal? Projects like Etoile and Beep Media Player use a wiki as their homepage and http://amarok.kde.org is mainly a front-end to it. And as I already said earlier, the spamming can be limited. Yes, there's the issue of defacement, but it won't happen that often if the software is kept up to date. Also, the wiki activity probably amounts to more than what the current homepage has seen recently. One alternative to having a wiki is having a well-maintained non-wiki-website, which is hard if you have got few people to do it. If I was Fedor or any other admin, I wouldn't just give anyone write-access. Another alternative would be having a CMS, but this wouldn't be much different, since you have to give people permission manually. I personally like the approach of Amarok. Have the main site be a bare-bones page with the most relevant stuff on it and link to the wiki for everything else. autogsdoc and texinfo remain the recommended ways to create manuals, but html (however created) would also be fine for linking from www.gnustep.org. If you can get it published in the community press, then that also helps to introduce new people to GNUstep. Please email link suggestions to [EMAIL PROTECTED] I wasn't talking about autogsdoc. I'm not totally against guides written in texinfo, but then you as the writer have the problem of maintaining it. New people are turned off by old and/or incomplete information. Besides, texinfo isn't the best format for linking to relevant information somewhere else, something a wiki does [[better]]. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
Am 06.09.2006 um 06:15 schrieb Rogelio M. Serrano Jr.: if we could not agree where scrollbars should be then scrollbars are bad and i would rather not have them. same goes for the menubar. While I agree with your wishes for a lean GUI, you are oversimplifying here. While some people think a car's steering wheel has to be on the left side and others prefer it on the right side, most of them agree you have to have it at least in either place. Markus - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 9/6/06, Markus Hitter [EMAIL PROTECTED] wrote: Am 06.09.2006 um 06:15 schrieb Rogelio M. Serrano Jr.: if we could not agree where scrollbars should be then scrollbars are bad and i would rather not have them. same goes for the menubar. While I agree with your wishes for a lean GUI, you are oversimplifying here. While some people think a car's steering wheel has to be on the left side and others prefer it on the right side, most of them agree you have to have it at least in either place. you are not going anywhere in a car without a steering wheel right? and the steering wheel is in the same location in relation to you, front and center. but then there is no configuration in a car right? somebody decided and you cant do anything about it. lets consider the stick gear shift. would you rather have stick to the left or right? or push buttons or rocker switch on the steering wheel with UPSHIFT and DOWNSHIFT labels where your thumb can get to it? ignore the the scroll bar in a browser. use the arrow keys or the search function and you got the same thing. i prefer the keys actually. -- the thing i like with my linux pc is that i can sum up my complaints in 5 items ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
On 4 Sep 2006, at 04:49, phil taylor wrote: For a software developer there is no substitute for hands on access to actual program code to get you up to speed on a new language. Why cant a GNUstep developer create even a simple hello world program and stick it in a samples directory. That might encourage new users to give the platform a look. I know it would work for me. GNUstep comes with a whole package of examples (see the downloads area of the website) in addition to the sample tools that come with the base library package. You can also use a load of tutorials and other documentation clearly linked from the web site. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On Wed, 2006-09-06 at 12:15 +0800, Rogelio M. Serrano Jr. wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Phil Taylor wrote: We are talking about GUI's on Linux, not GUI's in general. The Windows GUI is popular because Windows is popular and you cant have one without the other. windows is forced on you. and it has been around so long that people think thats thats how computers are supposed to work. Linux is different - you have the choice of GUI's, and both KDE and Gnome are already established and are highly configureable. People are used to having these facilities, and generally people dont like giving things up. I am sure that if someone were to release an configureable add on GUI for Windows (assuming it were technically possible which it is not) it would get a lot of attention. But then Windows users have never had that facility, so they dont miss it. yes thats right. thats why few people make a mess of their windows desktop configuration. if windows made it configurable it will be worse. its very easy to get x configuration wrong and not get a desktop at all. and its not hard to imagine how bad windows would be if it is like that. thats the same reason why the mac have limited user installable parts. and even less configurability in the gui. sure you can change themes but you cant change the basic gui elements. like have the scrollbars and menubars in different locations from one computer to the next. less configurability is so much better. its when you have bad design that you need customisations. I think we are getting confused between asthetics and bad design. The look and feel is highly subjective. There is no overall bad or good but simply what appeals to an individual. Anymore than there is absolute good or bad room decor. I have a room in my house painted dark green - lots of people would hate that colour scheme, but i like it. I am very glad that Gnome is themable, because the default theme sucks - its like GNUstep, very dark and dingy. Business like but there is no fun or lightness in it. Its like the difference between concrete and a garden. Windows default colur scheme sucks to, but you can change it a little. I do think (and this might be actually where you are coming from) that GUIs can be too configurable. I am quite happy accepting where the designer put the scroll bars in the all the major GUIs including GNUstep. But I like to be able to change the colour/pattern of the scroll bar. if we could not agree where scrollbars should be then scrollbars are bad and i would rather not have them. same goes for the menubar. So i stand by what I have said. Gnome and KDE are established and highly configurable. If GNUstep si to compete or even displace these GUI's it has to be percieved to be at least as up to date as they are. GUI's of no thats not going to work. If you mean that GNUstep wont replace either of these GUIs, i agree. It seems highly unlikely. That being said, I'm glad that themes and different menu styles are available in GNUstep and I think the default theme for GNUstep should be Nesedah. I am surprised that this is the case, since you never get to see any evidence of it. Why not make it more obvious? For example every time Gnome or KDE release a new version they supply new splash screens, colour schemes and themes. It cant take that much developer resources to at least change a few colours. yeah and you know how much effort is wasted on getting these things to work? M$ and Apple has big teams and the linux desktop people cant match that. and the more the linux desktop tries to compete on the same terms with windows and mac it will never be able to catch up. new ways need to be found. step out of the box people. WIMP is not the future. Surely putting in a new desktop background image or splash screen image doesnt require 5 man years of testing? I can change the background of my Gnome desktop in about 2 seconds, and its never failed yet. When you design an interface, you cannot assume that what you like someone else will like. Perhaps thats obvious. Well, it's more than just what somebody will like. In the case of Apple and NeXT, significant psychological research went into studying the way humans read, write, and interact with computer interfaces. As an example, we read from left-to-right in Western cultures and this effects our priority when performing a visual scan of an object. That's a fact that cannot be denied and has nothing to do with subjective appraisal of an interface. I guess all that research is irrelevant from the perspective of any particular user, such as myself (or you). I know what I like and what I am comfortable with. I basically dont give a rats arse that some researcher in Cupertino thinks the scroll bar should be on the left if i hah! until M$ decides its should be on the left! im sure the entire world
Re: scrollbars [was: Re: really attracting developers]
On 9/7/06, phil taylor [EMAIL PROTECTED] wrote: really? it momentum thats sustaining it. and people who is unwilling to find something new. I think you will find that the average user, who is not a hacker or an IT developer, will not find it acceptable to have to search for an i dont make that distinction among users. users like easy interfaces. easier the better. application once you have installed it. Indeed it is NOT acceptable in any sense of the word. And developers of Linux should hand their head in shame that they have made such an almighty bodge up of installing software. Probably the arrogance of software developers is what prevents them admitting this and doing something to change it. i would rather not have package management. i don't want to see applications at all. maybe have some kind of table with radio buttons for default application choices would be nice but the apps need to follow some kind of standard. If i had been even in part responsible for the design of the Unix directory structure and the system that is laughably referred oackage management (i think package mis-management is nearer the mark) i would conceal the fact from everyone and go hide somewhere. THis aspect of Linux/Unix is appaling and anyone who says different is deluded or seriously lacking in IQ. i don't like folders and directories. its ok for the system but tough on the users. flat filesystems are not a solution either. i have a solution but im not talking about it. an application menu still is not good enough for me but nextstep has a very nice and elegant solution to that. beats the windows start button. or the vista sidebar. i think first time users will like it. there is a significant number of people forced to use windows who needs better solutions and IT people including are ignoring them. Agreed. Windows sucks in a lot of areas. Also its made by Microsoft. QED. But then Linux aint so great in a lot of areas either. i was not making that kind of comparison. there is some sort of anti ordinary user culture in the whole IT industry. its the RTFM culture. you cant figure this out? READ THE FUCKING MANUAL!!! yeah good luck with that. if my grandmother cant figure it out by herself without reading a manual then im going back to the drawing board. if people would rather have pen paper then something is wrong. -- the thing i like with my linux pc is that i can sum up my complaints in 5 items ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 2006-09-04 19:01:33 +0200 Andrew Sveikauskas [EMAIL PROTECTED] wrote: [...] 3. An option to not show the app icon. [...] This is already implemented. You can set GSSuppressAppIcon = *BY; either on a per-application to suppress the application's miniwindow OR in NSGlobalDomain to supress all application miniwindows. -- Chris ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 5 Sep 2006, at 01:23, Andrew Ruder wrote: On Tue, Sep 05, 2006 at 09:34:04AM +1000, phil taylor wrote: I am dissapointed that the GNUstep project is devoted to its UI design. I had hoped the most important aspect was the API, not the look and feel of the GUI. IT will never suit me. I HATE menus, especially cascading ones. This is an argument I've started with many people many times on the IRC channel. These days the UI style of GNUstep actually detracts from people wanting to use it. While many people look at conforming with other toolkits look/feel as sacrificing features to become like everything else, I agree with you here. The API and programming language (Objective-C) are really what would attract many developers to GNUstep. Making GNUstep look/feel like other toolkits doesn't sacrifice anything important, IMO. And don't get me started on app bundles and other filesystem related things. I understand the benefits of these features, but as far as gaining popularity, they hurt more than they help. First ... Once we get theming working the appearance of the UI is largely IRRELEVANT. Second ... You may be right that the UI detracts from people wanting to use GNUstep. Certainly we get plenty of complaints about the look of the gui, but we have no basis to tell whether we get more complaints with the NeXT look and feel than we would if we showed another look and feel. Phil is just being making groundless assumptions when he complains that the GNUstep project is 'devoted to its UI design' ... since the project is actively moving forward on theming and few (no?) people argue against that, and while most of us probably like the NeXT UI (I haven't seen anything I like as much yet) that doesn't mean we aren't interested in trying new themes. Of course, we (most developers) put the API first, but clearly, even though Phil says 'I had hoped the most important aspect was the API' he immediately contradicts that with 'not the look and feel of the GUI. IT will never suit me'. I'd like to suggest that, when people write to the list complaining about their personal GUI tastes not being catered for, people refrain from commenting on the UI stuff altogether and simply direct those people to join in on adding/improving theming in the gui library. The perennial debate about what's good/bad in a user interface is, IMO, just a waste of everyones time. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 2006-09-05 13:52:25 +0200 Richard Frith-Macdonald [EMAIL PROTECTED] wrote: [...] The perennial debate about what's good/bad in a user interface is, IMO, just a waste of everyones time. Amen to that. Especially since any 5 people will have 6 or 7 different opinions on what really IS good or bad WRT interfaces... -- Chris ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 9/5/06, Chris Vetter [EMAIL PROTECTED] wrote: On 2006-09-05 13:52:25 +0200 Richard Frith-Macdonald [EMAIL PROTECTED] wrote: [...] The perennial debate about what's good/bad in a user interface is, IMO, just a waste of everyones time. Amen to that. Especially since any 5 people will have 6 or 7 different opinions on what really IS good or bad WRT interfaces... is that really workable? isnt the extreme configurability going to make the whole unwieldy? -- the thing i like with my linux pc is that i can sum up my complaints in 5 items ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Darwinports (was: Re: really attracting developers)
Wolfgang Keller wrote: Just a question: did you try to use: http://darwinports.opendarwin.org/ports/?by=namesubstr=GNUstep Yes. The portfile for GNUstep is broken. If this manifests itself with an error while gnustep-base is built, then the cause is likely that your darwinports tree and the darwinports programs are out of sync. The GNUstep portfiles make use of a port group file, which initially was installed by the gnustep-make package, but now seems to have moved into the base tree. So you either have the choice to update and reinstall the darwinports base tree or you back up the GNUstep group to a version which installs the necessary port group file. Skimming the change logs it looks like something like cvs update -dP -D06/01/2006 should do the job. Wolfgang ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
Am 05.09.2006 um 14:03 schrieb Rogelio Serrano: isnt the extreme configurability going to make the whole unwieldy? The whole X Windows and Motif system is configurable down into the smallest corner and most people live fine with this. The brilliant API and open source nature of GNUstep might help to get not yet again a resource hog. Markus - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: Darwinports (was: Re: really attracting developers)
CVS is not going to work anymore, because darwinports moved to macports.org, but port selfupdate should still work. Error: /opt/local/bin/port: selfupdate failed: Couldn't sync dports tree: sync failed doing rsync Sincerely, Wolfgang Keller -- My email-address is correct. Do NOT remove .nospam to reply. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
phil taylor [EMAIL PROTECTED] wrote: Please read the reasoning in this page before commenting further: http://www120.pair.com/mccarthy/nextstep/intro.htmld/ Please read the reasoning in this page before commenting further: - what a doozy of a comment. Everyone these days is a member of the gestapo. No, you asked what the rationale for the NeXT GUI was. Somebody thought it would be easier to redirect you to a site that explains in detail rather than regurgitating the whole thing. Of course being their project they can do what they like with it - its their perogative. But it would be nice at times to meet with developers who have more consideration for the people who may end up using their product, who are trying to please the end user rather than themselves. You've got to realize that this discussion has been had over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over for the past 5 years or so. And over that period of time, slow but steady progress has been made on changing the situtation. So when you come in here, like a charging bull, some people have an itchy trigger finger. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: Darwinports (was: Re: really attracting developers)
it should be working in /opt/local/etc/ports/sources.conf you should read rsync://rsync.darwinports.org/dpupdate/dports maybe the opendarwin servers were closed at some moments yves Le 06-09-05 à 15:34, Wolfgang Keller a écrit : CVS is not going to work anymore, because darwinports moved to macports.org, but port selfupdate should still work. Error: /opt/local/bin/port: selfupdate failed: Couldn't sync dports tree: sync failed doing rsync Sincerely, Wolfgang Keller -- My email-address is correct. Do NOT remove .nospam to reply. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
- Original Message - From: [EMAIL PROTECTED] To: phil taylor [EMAIL PROTECTED] Cc: discuss-gnustep@gnu.org Sent: Wednesday, September 06, 2006 5:22 AM Subject: Re: scrollbars [was: Re: really attracting developers] phil taylor [EMAIL PROTECTED] wrote: The many different ideas of what consitutes the ideal GUI, with what is silly to one person is sensible to another, shown on this mailing list, clearly seem to demonstrate that a GUI that is to appeal to a wide audience needs to be configurable. Anything less is destined only to find a small niche market at best. The Windows GUI is not configurable at all (you can change cosmetic things about it, but you cannot use vertical floating menus or a Mac style menu) and it dominates the market. The Mac GUI is not configurable either, really, and it dominates the little space that Windows leaves behind. So no, GUI configurability is not the key to success. The key to success is being in the right place at the right time. GNUstep has suffered from this more than anything else, as has the HURD. If GNUstep had been in its current state in 1996-1997, things would probably have been quite different. Likewise, if OPENSTEP/Mach had matured in the early 80s instead of the early 90s, it would very likely have been much more successful. We are talking about GUI's on Linux, not GUI's in general. The Windows GUI is popular because Windows is popular and you cant have one without the other. Linux is different - you have the choice of GUI's, and both KDE and Gnome are already established and are highly configureable. People are used to having these facilities, and generally people dont like giving things up. I am sure that if someone were to release an configureable add on GUI for Windows (assuming it were technically possible which it is not) it would get a lot of attention. But then Windows users have never had that facility, so they dont miss it. So i stand by what I have said. Gnome and KDE are established and highly configurable. If GNUstep si to compete or even displace these GUI's it has to be percieved to be at least as up to date as they are. GUI's of the past used to be battleship grey, and GNUstep looks distinctly as if from an earlier age - which of course it is, or at least the system it is replicating is. I seem to remember that NextStep ran on the Next machine which had monochrome graphics (at least initally), or perhaps it was colour but memory would have been limited? So an all grey scheme would have been acceptable in those times. That being said, I'm glad that themes and different menu styles are available in GNUstep and I think the default theme for GNUstep should be Nesedah. I am surprised that this is the case, since you never get to see any evidence of it. Why not make it more obvious? For example every time Gnome or KDE release a new version they supply new splash screens, colour schemes and themes. It cant take that much developer resources to at least change a few colours. When you design an interface, you cannot assume that what you like someone else will like. Perhaps thats obvious. Well, it's more than just what somebody will like. In the case of Apple and NeXT, significant psychological research went into studying the way humans read, write, and interact with computer interfaces. As an example, we read from left-to-right in Western cultures and this effects our priority when performing a visual scan of an object. That's a fact that cannot be denied and has nothing to do with subjective appraisal of an interface. I guess all that research is irrelevant from the perspective of any particular user, such as myself (or you). I know what I like and what I am comfortable with. I basically dont give a rats arse that some researcher in Cupertino thinks the scroll bar should be on the left if i want it on the right. And these issues are so subtle and complex that researchers are rarely able to come to an obvious and precise conclusion. Two companies will both employ such researchers, and then come out with different solutions. If this research yielded concrete and definitive results, then every GUI would look the same. no? But they dont. One thing that surprises me is that so much research has apparently been done on the GUI to improve user-computer interaction, and yet so little is done in other areas. Witness the appalingly dumb Unix directory structure, which is about as user friendly as a shark. A six year old kid could come up with a better structure than this. Or the idiotic situation with package installation, whereby under Unix you can install a package and then have no idea whatsoever how to get it to run, or even where the hell it is on the hard drive. e.g. you install a package from a web site advertising a product called DVD Scan, install it, and then stare at the blank desktop and wonder where is it and how do i run it. You scour the menus
Re: scrollbars [was: Re: really attracting developers]
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Phil Taylor wrote: We are talking about GUI's on Linux, not GUI's in general. The Windows GUI is popular because Windows is popular and you cant have one without the other. windows is forced on you. and it has been around so long that people think thats thats how computers are supposed to work. Linux is different - you have the choice of GUI's, and both KDE and Gnome are already established and are highly configureable. People are used to having these facilities, and generally people dont like giving things up. I am sure that if someone were to release an configureable add on GUI for Windows (assuming it were technically possible which it is not) it would get a lot of attention. But then Windows users have never had that facility, so they dont miss it. yes thats right. thats why few people make a mess of their windows desktop configuration. if windows made it configurable it will be worse. its very easy to get x configuration wrong and not get a desktop at all. and its not hard to imagine how bad windows would be if it is like that. thats the same reason why the mac have limited user installable parts. and even less configurability in the gui. sure you can change themes but you cant change the basic gui elements. like have the scrollbars and menubars in different locations from one computer to the next. less configurability is so much better. its when you have bad design that you need customisations. if we could not agree where scrollbars should be then scrollbars are bad and i would rather not have them. same goes for the menubar. So i stand by what I have said. Gnome and KDE are established and highly configurable. If GNUstep si to compete or even displace these GUI's it has to be percieved to be at least as up to date as they are. GUI's of no thats not going to work. That being said, I'm glad that themes and different menu styles are available in GNUstep and I think the default theme for GNUstep should be Nesedah. I am surprised that this is the case, since you never get to see any evidence of it. Why not make it more obvious? For example every time Gnome or KDE release a new version they supply new splash screens, colour schemes and themes. It cant take that much developer resources to at least change a few colours. yeah and you know how much effort is wasted on getting these things to work? M$ and Apple has big teams and the linux desktop people cant match that. and the more the linux desktop tries to compete on the same terms with windows and mac it will never be able to catch up. new ways need to be found. step out of the box people. WIMP is not the future. When you design an interface, you cannot assume that what you like someone else will like. Perhaps thats obvious. Well, it's more than just what somebody will like. In the case of Apple and NeXT, significant psychological research went into studying the way humans read, write, and interact with computer interfaces. As an example, we read from left-to-right in Western cultures and this effects our priority when performing a visual scan of an object. That's a fact that cannot be denied and has nothing to do with subjective appraisal of an interface. I guess all that research is irrelevant from the perspective of any particular user, such as myself (or you). I know what I like and what I am comfortable with. I basically dont give a rats arse that some researcher in Cupertino thinks the scroll bar should be on the left if i hah! until M$ decides its should be on the left! im sure the entire world will agree. like a flock of sheep! Its these reasons why Windows is still so popular. You can install software and find it afterwards. really? it momentum thats sustaining it. and people who is unwilling to find something new. there is a significant number of people forced to use windows who needs better solutions and IT people including are ignoring them. - -- things i hate about my linux pc: 1. it takes more than a second to boot up 2. keeps asking about filenames and directories 3. does not remember what i was working on yesterday 4. does not remember all changes i have ever made 5. cannot figure out necessary settings by itself -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFE/kt2yihxuQOYt8wRAkmeAJ4mqyYzw9o/hDiJjHPU8tNbJA4MjgCgi2rW GO5ijKDPQ51Kz2zgUIfKXgQ= =P3Ho -END PGP SIGNATURE- ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
Pete French wrote: to scroll down without a scrollwheel and without the keyboard I've to cross half of the screen to reach the scrollbar on the left side of the hmmm - whats your mouse doing on the right hand side for you to need to move it that far ? all the menus of a *step are on the left after all, so usually thats where your mouse is isn't it ? whats needed here is for people to desiign apps to that the controls are placed well - you'd get the same problem with right hand scrollbars if everything else was on the left. window. This is not nice for right handers. Its nice for left handers. this I do not understand either - it's the same amount of mouse movement no matter what hand you are holding the rodent with. I'm also right handed, but thats less important to me than the fact that I read text from left to right - so I want my scrollbars at the start of the line, and hence the left. I always wondered if the best solution would be for a text view to look at how the text is arranged and position the scroillbar accordingly (i.e. on the right if it's full of chinese or arabic, on the left for latin and cyrillic) but that might just be annoyingly inconsistent Actually I would love a NSGlobalDomain for that. well, configurability is always a good thing :-) I'm a developer so I spend almost 100% of my working day inside xterm, which has scrollabrs to the left and always has done as far as I know (for the usual reading text left to right reason as far as I know). am now off to see if theres a way to make it put them on the rght though, just for curiosities sake! :-) -bat. As Pete pointed it out correctly, the reason why the NeXT GUI is designed with left-hand scroll bars is because all important objects on the screen tend to aggregate on the upper left window margin. That's where the menu bar is even on M$ Windows. That's where all the toolbar buttons are on Thunderbird. That's where you start typing text in OpenOffice. Most of the user's focus on the left side of the window, which makes scrolling using left-hand scrollbars easier to track with peripheral viewing, rather than having to fly with the eyes across the entire window in order to move a document to a new offset. I agree with the default thing though - it would be very useful for right-to-left environments, where all of the user's focus is on the opposing side of the window. -- Saso ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
I'm also right handed, but thats less important to me than the fact that I read text from left to right - so I want my scrollbars at the start of the line, and hence the left. Oh. Never thought about that. This is a point for the left side indeed. I just observed that I always tend to move the mouse to the right if it hides some text. I want to have the mouse cursor go away, so I nudge the mouse away. And this is the right if you are right handed... :-) However I can life with that explanation. Thanks Marc ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
http://www120.pair.com/mccarthy/nextstep/intro.htmld/ After reading this something occured to me. NeXTstep is very nice, and via GNUstep I am well used to the way they do things. But a large quantity of people (most?) who are newly exposed to GNUstep are not looking to replace NeXTstep. This is probably what leads people to periodically complain on the mailing list. So, it seems the situation is like this: * Faction A loves GNUstep for its NeXT goodness * Faction B thinks GNUstep is out of place and should play nice with other desktops. * Probably some people believe both are true. So, my thinking was, why not offer a few NSUserDefaults to appease B above? This would include: 1. An option to have NSMenus appear within a window. It would mix better with an existing X or Win32 desktop and would also help solve the GNUstep doesn't work with focus follows mouse problem. 2. An option that makes all NSPanels visible regardless of what application has focus. This would solve the other half of the GNUstep doesn't work with focus follows mouse problem. 3. An option to not show the app icon. These three options alone would probably make some people complain less. But then I realized that, options to tweak the UI already exist, yet people still complain about the lack of Mac-like menus, etc. Maybe said people do not read documentation, or maybe they are not well documented, but, it does raise an important point: there needs to be a very clear, intuitive, idiot-proof way for new users to change UI styles. So my thought was very simple. It might be nice if gnustep-gui, upon running an application for the first time, popped up a panel that asked the user what kind of interface style they would want. There they could click away (select NeXT Mac or Win style menus, don't display the app icon, etc.) instead of being completely turned off by a program that doesn't fit their WM or desktop. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 2006-09-04 13:01:33 -0400 Andrew Sveikauskas [EMAIL PROTECTED] wrote: So, it seems the situation is like this: * Faction A loves GNUstep for its NeXT goodness * Faction B thinks GNUstep is out of place and should play nice with other desktops. * Probably some people believe both are true. True. So, my thinking was, why not offer a few NSUserDefaults to appease B above? This would include: Agreed. 1. An option to have NSMenus appear within a window. It would mix better with an existing X or Win32 desktop and would also help solve the GNUstep doesn't work with focus follows mouse problem. We will need to fake a main window when no documents are opened to do that. 2. An option that makes all NSPanels visible regardless of what application has focus. This would solve the other half of the GNUstep doesn't work with focus follows mouse problem. The problem with this is that it will increase screen clutter. 3. An option to not show the app icon. Agreed. These three options alone would probably make some people complain less. But then I realized that, options to tweak the UI already exist, yet people still complain about the lack of Mac-like menus, etc. Maybe said people do not read documentation, or maybe they are not well documented, but, it does raise an important point: there needs to be a very clear, intuitive, idiot-proof way for new users to change UI styles. Things will improve once WildMenus and theming is integrated into GUI. So my thought was very simple. It might be nice if gnustep-gui, upon running an application for the first time, popped up a panel that asked the user what kind of interface style they would want. There they could click away (select NeXT Mac or Win style menus, don't display the app icon, etc.) instead of being completely turned off by a program that doesn't fit their WM or desktop. Agreed, KDE does that and it is very nice. IMHO, there should some predefined themes that fixs in with the underlying system. Charles -- Never make any mistaeks. (Anonymous, in a mail discussion about to a kernel bug report.) RFC3156 defines security multipart formats for MIME with OpenPGP. pgpnn3qdrDQpx.pgp Description: PGP signature ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 2006-09-04 13:34:40 -0400 Charles Philip Chan [EMAIL PROTECTED] wrote: 2. An option that makes all NSPanels visible regardless of what application has focus. This would solve the other half of the GNUstep doesn't work with focus follows mouse problem. The problem with this is that it will increase screen clutter. Unfortunately if your WM is set for focus follows mouse or sloppy focus, the current behavior makes NSPanels almost unreachable, as they will disappear before your mouse can even get to them. When I started using GNUstep this bit me a lot, and I had to change my setting, which took some getting used to. The NeXT-style and Apple-style menus also do not get along with focus follows mouse. When you move the mouse outside the application the menu disappears. Seeing as this kind of focus is common on X desktops, I think fixing these 2 things would go a long way for some. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
I also think the web site could do much better to highlight these resources. How about a developers intro page, with sample code actually on the web? Allow me to appropriate your statement to advertise the wiki to possible contributors: http://mediawiki.gnustep.org/ It has a nicely visible link to developer guides. The thing is, I need help linking all the guides/tools listed on http://www.gnustep.org/developers/documentation.html and http://www.gnustep.org/experience/DeveloperTools.html. The wiki should eventually replace the current main page (www.gnustep.org) and *porting* all articles to the wiki would help this. BTW, new guides should be started on the wiki, not in some texinfo file :O That's all. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Just a question: did you try to use: http://darwinports.opendarwin.org/ports/?by=namesubstr=GNUstep Yes. The portfile for GNUstep is broken. Like, unfortunately, so many in Darwinports. Even the basic standards don't install for me. While in Fink, everything I've tried has worked so far. But GNUstep isn't available in Fink. Sincerely, Wolfgang Keller -- My email-address is correct. Do NOT remove .nospam to reply. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On Mon, 2006-09-04 at 12:26 +0100, Pete French wrote: to scroll down without a scrollwheel and without the keyboard I've to cross half of the screen to reach the scrollbar on the left side of the Can anyone (try to) explain the merits of the floating menus? I fail to see any advantages over the more usual menus which are either attached at some point to the application windows, or can be invoked by clicking anywhere within the window. The menus of multiple applications all look the same - only the title bar of the base menu distinguishes them, so with lots of apps open it gets harder and harder to locate which menu goes with which app. For every app open you now have two windows (main app window and menu window) open instead of one. So now you have two windows to position. Is that an advantage? I hardly see how anyone can believe that the dangling menu looks better. I suppose it looks odd to me largely due to it being different to the usual paradigm with which I am familiar, but even taking that into account it is still rather odd. As to the scrollbars being on the left, I have no basic objection to that, except to say that it makes sense to have the scroll bars on the same side as the close button, since those two operations happen most frequently. You open the window, scroll the text to locate something, then close it. If the close button is on the right and the scroll bar on the left, then that would indicate extra mouse movements are required in that situation. Overall, to me the GNUstep (and by defenition Openstep) interface seems to me odd for no good reason, as if it was dreamed up by marketing executives who see the need to differentiate a product against its competitors. I get the impression that GNUstep developers feel that if the made the GUI behave and look similar to KDE or GNOME, people would ignore the underlying architecture completely, and that their last hope of convincing people to use (GNUstep) it will fade. At least if it looks different people will think it IS different. Perhaps I should now go book myself in with a good psychiatrist! Phil T. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On Mon, 2006-09-04 at 13:01 -0400, Andrew Sveikauskas wrote: http://www120.pair.com/mccarthy/nextstep/intro.htmld/ After reading this something occured to me. NeXTstep is very nice, and via GNUstep I am well used to the way they do things. But a large quantity of people (most?) who are newly exposed to GNUstep are not looking to replace NeXTstep. This is probably what leads people to periodically complain on the mailing list. So, it seems the situation is like this: * Faction A loves GNUstep for its NeXT goodness * Faction B thinks GNUstep is out of place and should play nice with other desktops. * Probably some people believe both are true. So, my thinking was, why not offer a few NSUserDefaults to appease B above? This would include: 1. An option to have NSMenus appear within a window. It would mix better with an existing X or Win32 desktop and would also help solve the GNUstep doesn't work with focus follows mouse problem. 2. An option that makes all NSPanels visible regardless of what application has focus. This would solve the other half of the GNUstep doesn't work with focus follows mouse problem. 3. An option to not show the app icon. These three options alone would probably make some people complain less. But then I realized that, options to tweak the UI already exist, yet people still complain about the lack of Mac-like menus, etc. Maybe said people do not read documentation, or maybe they are not well documented, but, it does raise an important point: there needs to be a very clear, intuitive, idiot-proof way for new users to change UI styles. So my thought was very simple. It might be nice if gnustep-gui, upon running an application for the first time, popped up a panel that asked the user what kind of interface style they would want. There they could click away (select NeXT Mac or Win style menus, don't display the app icon, etc.) instead of being completely turned off by a program that doesn't fit their WM or desktop. YES! YES! YES!! Most if not all developers of other GUI s work on the assumption that people like to vary the appearance of their desktops - even Micro$oft. Also the primary users of GNUstep will of course be developers, but sually apps are developed to be used by end users, and not all of them may like floating menus and the Next look and feel. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 9/4/06, phil taylor [EMAIL PROTECTED] wrote: On Mon, 2006-09-04 at 12:26 +0100, Pete French wrote: to scroll down without a scrollwheel and without the keyboard I've to cross half of the screen to reach the scrollbar on the left side of the Can anyone (try to) explain the merits of the floating menus? I fail to see any advantages over the more usual menus which are either attached at some point to the application windows, or can be invoked by clicking anywhere within the window. The menus of multiple applications all look the same - only the title bar of the base menu distinguishes them, so with lots of apps open it gets harder and harder to locate which menu goes with which app. For every app open you now have two windows (main app window and menu window) open instead of one. So now you have two windows to position. Is that an advantage? I hardly see how anyone can believe that the dangling menu looks better. I suppose it looks odd to me largely due to it being different to the usual paradigm with which I am familiar, but even taking that into account it is still rather odd. The menu is auto-hiden. So even if you have a lot of GNUstep applications running, only the focused one has menu displayed. Some people have preference to put their menus. On mac, menus are stuck on the top. If you really like right-hand scrollbar, you can put the menu on the right side of your window, which may be more convenient to you. That's the advantage you are looking for. As to the scrollbars being on the left, I have no basic objection to that, except to say that it makes sense to have the scroll bars on the same side as the close button, since those two operations happen most frequently. You open the window, scroll the text to locate something, then close it. If the close button is on the right and the scroll bar on the left, then that would indicate extra mouse movements are required in that situation. How about scrolling the text to locate something and start to type ? With scrollbar on the left, your hand moves less to the position to start typing. Every UI design has a reason behind it. Please try to understand it first. Yen-Ju Overall, to me the GNUstep (and by defenition Openstep) interface seems to me odd for no good reason, as if it was dreamed up by marketing executives who see the need to differentiate a product against its competitors. I get the impression that GNUstep developers feel that if the made the GUI behave and look similar to KDE or GNOME, people would ignore the underlying architecture completely, and that their last hope of convincing people to use (GNUstep) it will fade. At least if it looks different people will think it IS different. Perhaps I should now go book myself in with a good psychiatrist! Phil T. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On Sep 5, 2006, at 24:38, phil taylor wrote: Can anyone (try to) explain the merits of the floating menus? You just need a single click and no drag to perform an arbitary (menu) action. You can easily detach the menu groups you need and thereby form some kind of favorite menus. A floating menu is more like a toolbar (or many toolbars) though a toolbar wastes space in every single window (but has better locality in return). I think there are no absolute pros or cons for either of the approaches. Both have their merits. If you have ever used NeXTstep you will know that vertical menus are very nice, a lookfeel is something hard to describe in words. Greets, Helge -- Helge Hess http://docs.opengroupware.org/Members/helge/ ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On Tue, Sep 05, 2006 at 09:34:04AM +1000, phil taylor wrote: I am dissapointed that the GNUstep project is devoted to its UI design. I had hoped the most important aspect was the API, not the look and feel of the GUI. IT will never suit me. I HATE menus, especially cascading ones. This is an argument I've started with many people many times on the IRC channel. These days the UI style of GNUstep actually detracts from people wanting to use it. While many people look at conforming with other toolkits look/feel as sacrificing features to become like everything else, I agree with you here. The API and programming language (Objective-C) are really what would attract many developers to GNUstep. Making GNUstep look/feel like other toolkits doesn't sacrifice anything important, IMO. And don't get me started on app bundles and other filesystem related things. I understand the benefits of these features, but as far as gaining popularity, they hurt more than they help. Dons-flame-retardant-cloak, Andy -- Andrew Ruder [EMAIL PROTECTED] http://www.aeruder.net ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
Am 05.09.2006 um 01:06 schrieb Pascal Bourguignon: On NeXTSTEP, I just kept the menu out of screen, and configured the right button to pop it up under the mouse.[...] Clearly, a menu bar is silly. Clearly, menus popping up where ever your mouse is, are silly. The human eye and brain remembers locations, and if the locations of menu items change all the time, it's extra work to pick them each time. No, I don't think GNUstep can find a general solution for all these personal preferences and I try to ignore these fruitless discussions, but I couldn't resist on this one. Markus - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 9/5/06, phil taylor [EMAIL PROTECTED] wrote: On Mon, 2006-09-04 at 12:26 +0100, Pete French wrote: to scroll down without a scrollwheel and without the keyboard I've to cross half of the screen to reach the scrollbar on the left side of the Can anyone (try to) explain the merits of the floating menus? I fail to see any advantages over the more usual menus which are either attached at some point to the application windows, or can be invoked by clicking anywhere within the window. i would rather not have a menu. -- the thing i like with my linux pc is that i can sum up my complaints in 5 items ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On Mon, 2006-09-04 at 18:58 -0400, Charles Philip Chan wrote: On 2006-09-04 18:46:26 -0400 phil taylor [EMAIL PROTECTED] wrote: Most if not all developers of other GUI s work on the assumption that people like to vary the appearance of their desktops - even Micro$oft. Skin deep appearance- yes, but pray tell how can I use a left scrollbar or menu palette in KDE, Gnome, Windows, or Aqua? Charles ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep True. But then it would be nice if you could, would'nt it? Appearance is not everything, or the most important thing, but it is not entirely irrelevant either. After all, command line interfaces are in many respects far more functional than GUI's, the best of which are still quite cumbersome. I guess if you put functionality right at the top of the list, you wouldnt use a GUI in the first place. But I take your point that the deabte is not purely about appearance. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On Tue, 2006-09-05 at 01:06 +0200, Pascal Bourguignon wrote: phil taylor writes: I hardly see how anyone can believe that the dangling menu looks better. I suppose it looks odd to me largely due to it being different to the usual paradigm with which I am familiar, but even taking that into account it is still rather odd. Just go to your favorite retaurant, and ask for the menu. Is it something like: Hors d'oeuvre Salade Waldorf Oeufs mimosa Plat Boeuf Bourguignon Escalope de veau Côtelettes d'agneau Truite aux amandes Desert Banana split Forêt noire Pomme or is it something like: Salade Waldorf, Oeufs mimosa, Boeuf Bourguignon, Escalope de veau, Côtelettes d'agneau, Truite aux amandes, Banana split, Forêt noire, Pomme But of course, I'll concede that trying to map real life stuff onto the computer screen is just a silly gimmick to sell more Mac to newbies, and that the best menu is the one that doesn't appear on the screen. Even the best window manager is the one that doesn't display anything on the screen (I'm going to leave WindowMaker for something like ratpoison or stumpwm soon). On NeXTSTEP, I just kept the menu out of screen, and configured the right button to pop it up under the mouse. For popup menus of course, the vertical disposition is better, because it needs less mouse movement to select the item. You must compare: - right button down, move a little down, right button up, vs. - move far, far, even farther nowadays with 1600x1200 displays, up left, (are you aready there?), right button down, move right, further, further, move down a little, right button up, move down, down, down back to the original place. Clearly, a menu bar is silly. And menu bars per window are worse. Good points. The thnig I like least about menus, especially the non pictorial ones like GNUstep (is that mandatory?) is that to find an item you have to look through the list - its harder to differentiate quickly. Pictures are much easier to locate. Toolbar buttons are expecially easy to locate because of the images - this is one of the founding original prinicples of the gui - icons. The main strength of menus is cramming lots of options in a small space. What you gain on the swings you lose on the roundabouts. I particularly hate cascading menus because of the tendency of the subordinate menu to dissappear if you do not move the mouse precisely enough, especially on high res screens with small text. Also any hierarchical setup makes finding the subordinate items very hard, as it is very difficult to traverse a complex tree and remember where you have already been. Thats another pet hate of mine - the awful Unix directory structure, with its /lib/bin/share/bin/lib structure. e.g. Is it in /lib/bin/hotplug or /lib/hotplug/bin or /lib/sbin/local/share/hotplug or lib/hotplug/local/bin/share etc... I think you get the point. Probably i am too impatient and menus are best when used sedately. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On Mon, 2006-09-04 at 19:23 -0500, Andrew Ruder wrote: On Tue, Sep 05, 2006 at 09:34:04AM +1000, phil taylor wrote: I am dissapointed that the GNUstep project is devoted to its UI design. I had hoped the most important aspect was the API, not the look and feel of the GUI. IT will never suit me. I HATE menus, especially cascading ones. This is an argument I've started with many people many times on the IRC channel. These days the UI style of GNUstep actually detracts from people wanting to use it. While many people look at conforming with other toolkits look/feel as sacrificing features to become like everything else, I agree with you here. The API and programming language (Objective-C) are really what would attract many developers to GNUstep. Making GNUstep look/feel like other toolkits doesn't sacrifice anything important, IMO. And don't get me started on app bundles and other filesystem related things. I understand the benefits of these features, but as far as gaining popularity, they hurt more than they help. Dons-flame-retardant-cloak, Andy Someone who agrees with me!!! I must be dreaming. I guess there is a first time for everything - honey, get the Guiness Book of Records on the phone! Have you perhaps been smoking something? You are not joking are you! Anyone offering advice or comments about a software project with the idea of trying to change its direction are at a serious disadvantage. People who run and work on these projects seem to get very defensive about their babies and would rather do anything that admit they may be wrong or that their idea of the worlds best feature is not really shared by many others. Of course being their project they can do what they like with it - its their perogative. But it would be nice at times to meet with developers who have more consideration for the people who may end up using their product, who are trying to please the end user rather than themselves. A little bit less of the this is what I have written and if you dont like it bugger off and a bit more of oh really, you would like that changed - well thanks for telling me, ill look into it. Sadly this less selfish attitude is seen much less often than the f'ait accompli approach most developers seem to have. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: Re: scrollbars [was: Re: really attracting developers]
On Tue, 2006-09-05 at 03:28 +0100, Nicolas Roard wrote: On 9/5/06, phil taylor [EMAIL PROTECTED] wrote: On Tue, 2006-09-05 at 00:51 +0200, Helge Hess wrote: On Sep 5, 2006, at 24:38, phil taylor wrote: Can anyone (try to) explain the merits of the floating menus? You just need a single click and no drag to perform an arbitary (menu) action. You can easily detach the menu groups you need and thereby form some kind of favorite menus. A floating menu is more like a toolbar (or many toolbars) though a toolbar wastes space in every single window (but has better locality in return). I think there are no absolute pros or cons for either of the approaches. Both have their merits. If you have ever used NeXTstep you will know that vertical menus are very nice, a lookfeel is something hard to describe in words. Greets, Helge Thanks. So the real advantage seems to come when you detach the lower level menus and personalise the interface. But doesnt that use up a lot of space? Do you have to do that each time the app is loaded? Of course not. One of the good things with GNUstep apps is that all this stuff is persistant -- I close an application, I reopen it, and all the opened submenus come back where they were, and the different windows reopen at the same place too. Persistance is good... :-) Secondly, the screen space they take is not very important -- first because you can obviously close them if you need, secondly because things that are not windows (ie, panels and menus) disappear automatically when you switch to another application. Imagine that you are using TextEdit. You opened the font panel so you can easily play with font settings. You also have 2-3 submenus opened, that you moved somewhere on the screen where it is more convenient for your current use. You then click on GNUMail -- automatically the TextEdit font panel and the menus disappear, the only TextEdit thing left on the screen is the actual window containing your text; and GNUMail's own panels or menus appears. That behaviour seriously help reducing the screen clutter. I am dissapointed that the GNUstep project is devoted to its UI design. I had hoped the most important aspect was the API, not the look and feel of the GUI. IT will never suit me. I HATE menus, especially cascading ones. Because you never used proper ones :-P just kidding. I agree with you -- the most important thing is probably the API, not the UI. But the whole UI experience is important too, and if something is good, it's a bit stupid to want to throw it by the window, don't you think ? Anyway, you do not have to use vertical menus if you don't like them. Just set the defaults NSInterfaceStyle to use horizontal menus, or use the EtoileMenu bundle. Personally, I like vertical menus, I think they work brilliantly with big screens. On the other hand, horizontal menus (ala mac) work better on small screens. Horizontal menus also have the added functionality that it's easy to add things like a system menu (eg the Apple menu on Mac OS) or menulet (clocks, virtual desktop, battery, user switching, sound, whatever) ; with vertical menus you can't really do that, so it means you need another place on the screen to put this kind of info/actions. That's why on étoilé we finally settled on using horizontal menu, like OS X. Probably also because it's much easier to convaince people to use horizontal menu than vertical menu, and fitt's law is with you. But still, I love vertical menus :-) and optional vertical menus is certainly a good thing. Is this étoilé project a fork of GNUstep? It looks interesting. What I would like is for the GNUstep api to be integrated with GTK+. Now that would be something i would go for. But obviously from my conversations on this mailing list, very unlikely to ever happen. You are not very helpful :-) You ask people to tell you what the UI is good for, as they seem to think it's not as crappy as you'd consider yourself; then you tell people that they value more the UI than the API (which they never said). Then the final conclusion is to integrate GNUstep api to GTK+ ? are you joking ? That can't happen, not for political reasons (well.. not only..) but simply because it wouldn't work -- philosophically and technically the two api are really different. In the end you'd need to redo/clone lots of work on either side, to end up with something that would be worse. Beside, a good reason OpenStep is that clean is because of Objective-C -- you'd need an OO language as dynamic as it. I didnt really mean literally to integrate GTK+ and GNUstep api. I meant to have the dynamic obj-c based GNUstep api but using a GUI paradigm that looked like GTK+. In other words I want to be able to created a GNUstep app that when you run it looks indistinguishable from an app written using Glade, GTK+ and the Gnome api's. Of
Re: [Fwd: Re: really attracting developers]
It already is released: http://wwwmain.gnustep.org/experience/examples.html (From the 'Applications' link on the main page) links to ftp://ftp.gnustep.org/pub/gnustep/core/gnustep-examples-1.0.0.tar.gz been there for about 2 years... On Sep 3, 2006, at 11:25 PM, Gregory John Casamento wrote: Phil, There is also an examples directory in the repository under usr-apps available in the SVN repository. We should probably start releasing it in a separate examples tarball to illustrate some of these things. Later, GJC --Gregory John Casamento - Original Message From: phil taylor [EMAIL PROTECTED] To: Gregory John Casamento [EMAIL PROTECTED] Cc: GNUstep Discussion discuss-gnustep@gnu.org Sent: Monday, September 4, 2006 12:37:46 AM Subject: Re: [Fwd: Re: really attracting developers] On Sun, 2006-09-03 at 21:08 -0700, Gregory John Casamento wrote: Phil, phil taylor [EMAIL PROTECTED]: If they cant bother to spend a couple of hours to knock up some sample programs, or build a few packages for the major distros (that arent four years out of date) what kind of message does that send to new users about the platform and the developers of it? I know the message it send to me - go somewhere else. In the Gorm release, since about 0.5.0 (most current release is 1.1) there are examples in the Documentation directory. They are covered in the Gorm manual which is also in that directory. [EMAIL PROTECTED] Examples]$ ls Controller SimpleApp [EMAIL PROTECTED] Examples]$ pwd /home/heron/Releases/gorm-1.1.0/Documentation/Examples [EMAIL PROTECTED] Examples]$ There are your simple examples. They've been there for a quite a while. Later, GJC And what if you dont know about Gorm or have decided not to use it? You might well want to try building a simple console app before a GUI one, or might want to build the GUI one by hand so as to get the feel of the code, rather than have it generated? I think the sample code should be included in the base development install, or perhaps with every dev tool? I also think the web site could do much better to highlight these resources. How about a developers intro page, with sample code actually on the web? I can find lots of other stuff on the site, but its the things that are essential for a first time user are not easy to find. And its the first time users who do not know either the software or the site that need these things. Experienced users have found the resources for themselves. This is the usual situation with open source projects - the info is all there, but you have to either be very lucky to stumble on it quickly, or spend months looking for a needle in a haystack. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: Re: scrollbars [was: Re: really attracting developers]
On 2006-09-05 00:35:58 -0400 phil taylor [EMAIL PROTECTED] wrote: Take the case of the side by side lists used to browse the directory structure within GNUstep. Its the most functional and easiest method I have ever come across for directory browsing. Yes, browser (column) view is great. But also simply gawky and unattractive to look at. Each to hir own. I prefer simple, non-obtusive GUI's. Do Apple still use this setup for directory browsing? Yes, but they took away the icon path which makes it rather useless. Charles -- I did this 'cause Linux gives me a woody. It doesn't generate revenue. (Dave '-ddt-` Taylor, announcing DOOM for Linux) RFC3156 defines security multipart formats for MIME with OpenPGP. pgpuZiADw8TdP.pgp Description: PGP signature ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On Tue, 2006-09-05 at 12:38 +0800, Rogelio Serrano wrote: On 9/5/06, phil taylor [EMAIL PROTECTED] wrote: On Tue, 2006-09-05 at 03:21 +0200, Markus Hitter wrote: Am 05.09.2006 um 01:06 schrieb Pascal Bourguignon: you have the source go ahead and create another ui. Im over fifty. Judging by how long its taking to get GNUstep out of the door, i think ill pass. Besides, if I did i would get all these jerks on the mailing list telling me that I had stuffed up the user interface - who wants to put up with that!!! ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 9/5/06, phil taylor [EMAIL PROTECTED] wrote: On Tue, 2006-09-05 at 12:38 +0800, Rogelio Serrano wrote: On 9/5/06, phil taylor [EMAIL PROTECTED] wrote: On Tue, 2006-09-05 at 03:21 +0200, Markus Hitter wrote: Am 05.09.2006 um 01:06 schrieb Pascal Bourguignon: you have the source go ahead and create another ui. Im over fifty. Judging by how long its taking to get GNUstep out of the door, i think ill pass. yeah WIMP ui's are complicated. its not friendly to small teams. so none-WIMP is the future for me. Besides, if I did i would get all these jerks on the mailing list telling me that I had stuffed up the user interface - who wants to put up with that!!! you are not going to force it on everybody else right? i have my ui you have your ui. i might diss yours and diss mine but hey thats life. -- the thing i like with my linux pc is that i can sum up my complaints in 5 items ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: scrollbars [was: Re: really attracting developers]
On 2006-09-05 00:47:45 -0400 phil taylor [EMAIL PROTECTED] wrote: Besides, if I did i would get all these jerks on the mailing list telling me that I had stuffed up the user interface - who wants to put up with that!!! Ha, ha. I have no problems discussing different GUI's, although I prefer NeXTstep. In fact collecting GUI's is one of my hobbies. To date, I have used Desqview/X, Windows (from version 3), RISC OS GUI, Desk Mate, GEM, Mac OS (starting from the version that came with an Apple II), OS/2, varies WM for Unix, etc. However, I have problems with people criticizing things that s/he haven't even attempted to try out. Charles -- /* Am I fucking pedantic or what? */ linux-2.2.16/drivers/scsi/qlogicpti.h RFC3156 defines security multipart formats for MIME with OpenPGP. pgplg8agG8orq.pgp Description: PGP signature ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Well, I will never know, Although I put like 100 hours in trying to compile gnustep on my intel mac, i didn't get even close of getting there. (I even tried installing plain darwin, but that could not fix the problems.) So it was like impossible to get it running, so the first step of porting an osx app to gnustep was already to hard. Then I thought of installing gnustep on my windows box instead. Installation was a dream, due to the great package provided for windows. But after that, gnustep did not even compile most of the apps that were available on the gnustep site itself. And also the developer-tools were not apparent how to use, and did even crash, or had unexpected behavior. The gnustep site did not mention it was developed for main usage on linux. So, after finding out, I installed linux, and used binary packages. Now most things did work. BUT I DON'T WANT TO USE LINUX! I want to use this great nextstep framework and the beauty of objective-c to work on *windows* Because I want my apps to be *available* like that is why I hate the mac cocoa framework, and gnustep would be it! And that is why I don't want to develop for linux! On windows, the gnustep framework needs to be installed first. And it does not look like any other windows app, so users don't understand the slightest bit what to do. *unlike* when using osx! This is not acceptable. Because I cannot expect them to figure it out, they aren't programmers. So I didn't even get to that great ide/libraries to test them in the first place... Now I figured out that objective-c and windows libraries don't even compile together, there is very little I can do with my OSX code on a pc... I really would like to develop for gnustep. But: - it has to run on windows smoothly and beautifully, with the app having a windows like gui. (Come on guys, windows is the most used platform! This would bring nextstep to the masses!) (- same maybe for linux/unix, with the app having a more common unix like experience (kde?) (Don't want to leave ppl out, right)) - Also remarks should be added to the website about best practices, so that not too many hours should be spend on setting up configurations that do not give the expected results. - Developer-tools should look and behave more finished - There should be converter tools to assist making osx source run on gnustep So this is why this developer is still kept from gnustep (unwillingly) Ewoud On Aug 25, 2006, at 12:05 PM, Rogelio Serrano wrote: Whats keeping other developers from gnustep? incomplete ide? incomplete nextstep based system? incomplete libraries? i don't buy the general applications unavailability argument. we are talking about people who want to create apps under gnustep. i don't buy the appearance argument either. -- things i hate about my linux pc: 1. it takes more than a second to boot up 2. keeps asking about filenames and directories 3. does not remember what i was working on yesterday 4. does not remember all the changes i have ever made 5.cannot figure out necessary settings by itself ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ** Scanned by VisNetic MailScan for SMTP Servers. Visit http://www.deerfield.com/products/visnetic_mailscan. ** ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 9/3/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: first place... Now I figured out that objective-c and windows libraries don't even compile together, there is very little I can do with my OSX code on a pc... I really would like to develop for gnustep. But: - it has to run on windows smoothly and beautifully, with the app having a windows like gui. (Come on guys, windows is the most used platform! This would bring nextstep to the masses!) does that matter? seriously? if thats important for me i would not be using open source at all. what is it exactly that you like about nextstep? the language? i dont really understand what you are trying to do. you are a windows programmer right? you would have to do that yourself or you could pay someone else to do that for you. but i doubt if a microsoft certified programmer would like to work with gnustep. nextstep is all about the vertical menus and left hand vertical scroll bar. its a complete ui design and it does not fit in windows. make it behave like windows and you just reinvented windows. who wants gnumail with a horizontal menubar and right hand vertical scrollbar? or have ribbon instead like office 2007? -- the thing i like with my linux pc is that i can sum up my complaints in 5 items ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On Sep 3, 2006, at 1:16 PM, Rogelio Serrano wrote: having a windows like gui. (Come on guys, windows is the most used platform! This would bring nextstep to the masses!) does that matter? seriously? if thats important for me i would not be using open source at all. So as an open source developer, you don't care about recognition, or wide usability of your software...? And who said that I would be only developing open source? Why not developing commercial software, and help develop gnustep at the same time, because I need some specific functionality. And some open- source for the fun of it. what is it exactly that you like about nextstep? the language? i dont really understand what you are trying to do. I want to use it's api. Like nextstep an cocoa, it develops programs in a *beautiful* *fast* *simple* and *gratifying* way. More gratifying than programs written in any other framework. you are a windows programmer right? you would have to do that yourself or you could pay someone else to do that for you. but i doubt if a microsoft certified programmer would like to work with gnustep. Well I have a Microsoft background. And I am sufficiently fed up with it.. So I was looking to writing osx, but writing for osx means having such a limited audience. They are so *protective*, which is really discouraging. I am fully prepared to make a total new start. So I stared looking into gnustep. Like I already said, I could not get it right. So now I am stuck with writing for osx. nextstep is all about the vertical menus and left hand vertical scroll bar. its a complete ui design and it does not fit in windows. make it behave like windows and you just reinvented windows. I am sorry that I want to produce something practical... (I am using linux by the way, as a firewall. I hope I did not offend you personally.) Ewoud ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Am 03.09.2006 um 12:36 schrieb [EMAIL PROTECTED]: Well, I will never know, Although I put like 100 hours in trying to compile gnustep on my intel mac, i didn't get even close of getting there. (I even tried installing plain darwin, but that could not fix the problems.) Just a question: did you try to use: http://darwinports.opendarwin.org/ports/?by=namesubstr=GNUstep or did you follow: http://www.gnustep.org/resources/documentation/User/GNUstep/ README.Darwin (a little bit out of date I fear. There was a wikipage derived from that but I don't know where this is gone ... ) Some additional info: the DarwinPorts project is in the process of migrating to http:// www.macports.org/, so expect some rough edges. But you can always ask for help at: [EMAIL PROTECTED] especially look out for a guy called Yves de Champlain ([EMAIL PROTECTED] darwin.org) So it was like impossible to get it running, so the first step of porting an osx app to gnustep was already to hard. Then I thought of installing gnustep on my windows box instead. Installation was a dream, due to the great package provided for windows. But after that, gnustep did not even compile most of the apps that were available on the gnustep site itself. And also the developer-tools were not apparent how to use, and did even crash, or had unexpected behavior. The gnustep site did not mention it was developed for main usage on linux. So, after finding out, I installed linux, and used binary packages. Now most things did work. BUT I DON'T WANT TO USE LINUX! I want to use this great nextstep framework and the beauty of objective-c to work on *windows* Because I want my apps to be *available* like that is why I hate the mac cocoa framework, and gnustep would be it! And that is why I don't want to develop for linux! On windows, the gnustep framework needs to be installed first. And it does not look like any other windows app, so users don't understand the slightest bit what to do. *unlike* when using osx! This is not acceptable. Because I cannot expect them to figure it out, they aren't programmers. So I didn't even get to that great ide/libraries to test them in the first place... Now I figured out that objective-c and windows libraries don't even compile together, there is very little I can do with my OSX code on a pc... I really would like to develop for gnustep. But: - it has to run on windows smoothly and beautifully, with the app having a windows like gui. (Come on guys, windows is the most used platform! This would bring nextstep to the masses!) (- same maybe for linux/unix, with the app having a more common unix like experience (kde?) (Don't want to leave ppl out, right)) - Also remarks should be added to the website about best practices, so that not too many hours should be spend on setting up configurations that do not give the expected results. - Developer-tools should look and behave more finished - There should be converter tools to assist making osx source run on gnustep So this is why this developer is still kept from gnustep (unwillingly) Ewoud On Aug 25, 2006, at 12:05 PM, Rogelio Serrano wrote: Whats keeping other developers from gnustep? incomplete ide? incomplete nextstep based system? incomplete libraries? i don't buy the general applications unavailability argument. we are talking about people who want to create apps under gnustep. i don't buy the appearance argument either. -- things i hate about my linux pc: 1. it takes more than a second to boot up 2. keeps asking about filenames and directories 3. does not remember what i was working on yesterday 4. does not remember all the changes i have ever made 5.cannot figure out necessary settings by itself ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ** Scanned by VisNetic MailScan for SMTP Servers. Visit http://www.deerfield.com/products/visnetic_mailscan. ** ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 9/3/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: On Sep 3, 2006, at 1:16 PM, Rogelio Serrano wrote: having a windows like gui. (Come on guys, windows is the most used platform! This would bring nextstep to the masses!) does that matter? seriously? if thats important for me i would not be using open source at all. So as an open source developer, you don't care about recognition, or wide usability of your software...? i do. i want to create useful programs. but i did not choose linux because it is popular. i chose it because i think it is did some things right and it is accessible without expensive fees. so i think open source is the way to bring nextstep to the masses. And who said that I would be only developing open source? Why not developing commercial software, and help develop gnustep at the same time, because I need some specific functionality. And some open- source for the fun of it. i have nothing against that. what is it exactly that you like about nextstep? the language? i dont really understand what you are trying to do. I want to use it's api. Like nextstep an cocoa, it develops programs in a *beautiful* *fast* *simple* and *gratifying* way. More gratifying than programs written in any other framework. well thats good. you are a windows programmer right? you would have to do that yourself or you could pay someone else to do that for you. but i doubt if a microsoft certified programmer would like to work with gnustep. Well I have a Microsoft background. And I am sufficiently fed up with it.. So I was looking to writing osx, but writing for osx means i see. so thats where you are coming from. welcome to the club. having such a limited audience. They are so *protective*, which is really discouraging. yes, been there... I am fully prepared to make a total new start. So I stared looking into gnustep. Like I already said, I could not get it right. So now I am stuck with writing for osx. i hope someone can help you. nextstep is all about the vertical menus and left hand vertical scroll bar. its a complete ui design and it does not fit in windows. make it behave like windows and you just reinvented windows. I am sorry that I want to produce something practical... (I am using linux by the way, as a firewall. I hope I did not offend you personally.) its ok. i use linux at work and at home. i am being paid to produce something practical too. -- the thing i like with my linux pc is that i can sum up my complaints in 5 items ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
I want to use this great nextstep framework and the beauty of objective-c to work on *windows* Out of interest, did you ever try using Apple's developent kit for Windows ? I know they discontinued it, but there are still copies floating around, and it rather nice - also gives you an idea of how close a *Step app can come to fitting in with Windows without a re-write. -bat. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Do you mean the Itunes/quicktime development kits, or do you mean rhapsody? Quicktime and Itunes are based on carbon, not on cocoa (the renamed openstep api) And if I am well informed the rhapsody license is expired. Nice if you want make apps for your own, but not so handy if you want to release something. You could not because you would use the framework, and so would also need a license to use and distribute it. I hope I am wrong with this, and if you have more suggestions, they would be very welcome. Ewoud On Sep 3, 2006, at 5:11 PM, Pete French wrote: I want to use this great nextstep framework and the beauty of objective-c to work on *windows* Out of interest, did you ever try using Apple's developent kit for Windows ? I know they discontinued it, but there are still copies floating around, and it rather nice - also gives you an idea of how close a *Step app can come to fitting in with Windows without a re-write. -bat. ** Scanned by VisNetic MailScan for SMTP Servers. Visit http://www.deerfield.com/products/visnetic_mailscan. ** ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Do you mean the Itunes/quicktime development kits, or do you mean rhapsody? I meant the YellowBox stuff and it's predescessor - but (as you say) it's no use for deploying to real users. It is, however, of interest 9to me anyway) as a way of seeing how OpenStep apps work in a Windows environment with that look. -bat. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On Sun, 2006-09-03 at 14:58 +0200, [EMAIL PROTECTED] wrote: nextstep is all about the vertical menus and left hand vertical scroll bar. its a complete ui design and it does not fit in windows. make it behave like windows and you just reinvented windows. Seriously? All that API wizardry, and the entire object oriented framework is just there so you can put the scroll bar on the left and have a detached menu? This is a very trivial assessment, and by definition total nonsense. Surely GNUStep is mainly about the application development framework - thats what GNUstep actually is - and how it enables one to develop applications very quickly and very reliably. Personally I think the detached menus and certain other aspects of the GUI, but especially the floating menus, are ugly and quite off putting. I cannot imagine why anyone would think they are a good idea. I would imagine the look of the GNUstep gui is one of the main reasons GNUstep is nowhere near as popular as it could be. The other main reason would have to be the lack of a simple and effective way to install it. I have just managed to get it working on Debian and am not very impressed. Even though there is a package in the Debian repos, not all the set up is done (environment path variables not configured). There appear to be not so much as a single sample program - goodness knows it wouldnt take a GNUstep developer more than a few minutes to knock up some simple sample programs. The documentation supplied is incomplete - just a ragged collection of bits and bobs. Not of much use to a newbie. Despite these other objections, I have decided agaisnt using it (so far) on account of Objective-C rather than GNUstep per se. The object features have an obvious bolted-on look to them. Class definition syntax is a real alphabet soup. The code looks really unreadable. The only idea I like is the named parameters, but the object structure itself does not look very elegant. Think I will go with SmartEiffel and GTK+. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Even though there is a package in the Debian repos, not all the set up is done (environment path variables not configured). Does Debian even have a way for a package to do this? Wouldn't it be much nicer if a Debian package _didn't_ mess with user-level shell rc files? There appear to be not so much as a single sample program - goodness knows it wouldnt take a GNUstep developer more than a few minutes to knock up some simple sample programs. This is rubbish. Just 3 clicks away from the home page on gnustep.org, I was able to find the GNUstep Examples package: ftp://ftp.gnustep.org/pub/gnustep/core/gnustep-examples-1.0.0.tar.gz ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 9/4/06, phil taylor [EMAIL PROTECTED] wrote: On Sun, 2006-09-03 at 14:58 +0200, [EMAIL PROTECTED] wrote: nextstep is all about the vertical menus and left hand vertical scroll bar. its a complete ui design and it does not fit in windows. make it behave like windows and you just reinvented windows. Seriously? All that API wizardry, and the entire object oriented framework is just there so you can put the scroll bar on the left and have a detached menu? This is a very trivial assessment, and by definition total nonsense. of course the api is included. nextstep is all or nothing for me. if you use the api and not the look thats not nextstep. its academic for me because im doing exactly the same thing. im even going beyond what the original poster is doing. im using gnustep to build a non wimp ui. no menu no overlapping windows no scrollbars no apps no folders no icons no window manager. I would imagine the look of the GNUstep gui is one of the main reasons GNUstep is nowhere near as popular as it could be. The other main reason would have to be the lack of a simple and effective way to install it. i dont agree. its not popular because it does not fit on the platforms we shoehorned it onto. I have just managed to get it working on Debian and am not very impressed. Even though there is a package in the Debian repos, not all the set up is done (environment path variables not configured). There appear to be not so much as a single sample program - goodness knows it wouldnt take a GNUstep developer more than a few minutes to knock up some simple sample programs. gnustep is supposed to be a development platform not a desktop. i take that as saying that you use native tools to configure the plaform. for me on debian that means lots of xterm windows alongside gnumail and gworkspace. i wonder where nextstep consistency went. Think I will go with SmartEiffel and GTK+. to each his own... i prefer dylan but well... next year maybe... -- the thing i like with my linux pc is that i can sum up my complaints in 5 items ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On Sep 2, 2006, at 3:26 AM, Nikolaus Waxweiler wrote: I don't see how Mediawiki would be WCAG-hostile, given that Wikipedia uses it. If you're referring to the captcha/logic puzzle test, this is just a minor hurdle. By the way, the automatically generated API documentation will not be moved to the wiki or deprecated. The one in the wiki will be used for code snippets and other things and will link to the generated one, incorporating the best of both worlds :) I definitely think something needs to change with the web site. The wiki gets updated more quickly because almost anyone can go in a change it. If I can get many people to do updates instead of one or two developers having to do all of it, I'd be more than happy. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
Phil, phil taylor [EMAIL PROTECTED]: If they cant bother to spend a couple of hours to knock up some sample programs, or build a few packages for the major distros (that arent four years out of date) what kind of message does that send to new users about the platform and the developers of it? I know the message it send to me - go somewhere else. In the Gorm release, since about 0.5.0 (most current release is 1.1) there are examples in the Documentation directory. They are covered in the Gorm manual which is also in that directory. [EMAIL PROTECTED] Examples]$ ls Controller SimpleApp [EMAIL PROTECTED] Examples]$ pwd /home/heron/Releases/gorm-1.1.0/Documentation/Examples [EMAIL PROTECTED] Examples]$ There are your simple examples. They've been there for a quite a while. Later, GJC --Gregory John Casamento - Original Message From: phil taylor [EMAIL PROTECTED] To: GNUstep Discussion discuss-gnustep@gnu.org Sent: Sunday, September 3, 2006 11:49:25 PM Subject: [Fwd: Re: really attracting developers] Subject: Re: really attracting developers From: phil taylor [EMAIL PROTECTED] To: Rogelio Serrano [EMAIL PROTECTED] In-Reply-To: [EMAIL PROTECTED] References: [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] Content-Type: text/plain Message-Id: [EMAIL PROTECTED] Mime-Version: 1.0 X-Mailer: Evolution 2.6.3 Date: Mon, 04 Sep 2006 13:48:20 +1000 Content-Transfer-Encoding: 7bit On Mon, 2006-09-04 at 11:12 +0800, Rogelio Serrano wrote: On 9/4/06, phil taylor [EMAIL PROTECTED] wrote: On Sun, 2006-09-03 at 14:58 +0200, [EMAIL PROTECTED] wrote: gnustep is supposed to be a development platform not a desktop. i take that as saying that you use native tools to configure the plaform. for me on debian that means lots of xterm windows alongside gnumail and gworkspace. i wonder where nextstep consistency went. Of course I am commenting on it from a developers perspective. What use would sample program code be to an end user? Go download Freebasic (from www.freebasic.org) and install it. You get a samples directory with numerous programs dealing with all aspects for application creation, e.g. using GTK, OpenGL, mail, file io, etc, in addition to the documentation. For a software developer there is no substitute for hands on access to actual program code to get you up to speed on a new language. Why cant a GNUstep developer create even a simple hello world program and stick it in a samples directory. That might encourage new users to give the platform a look. I know it would work for me. Aside from the benefit of the code itself, the absence of code speaks volumes about how anxious the developers are to get new people on board. If they cant bother to spend a couple of hours to knock up some sample programs, or build a few packages for the major distros (that arent four years out of date) what kind of message does that send to new users about the platform and the developers of it? I know the message it send to me - go somewhere else. next year maybe... ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
On Sun, 2006-09-03 at 21:08 -0700, Gregory John Casamento wrote: Phil, phil taylor [EMAIL PROTECTED]: If they cant bother to spend a couple of hours to knock up some sample programs, or build a few packages for the major distros (that arent four years out of date) what kind of message does that send to new users about the platform and the developers of it? I know the message it send to me - go somewhere else. In the Gorm release, since about 0.5.0 (most current release is 1.1) there are examples in the Documentation directory. They are covered in the Gorm manual which is also in that directory. [EMAIL PROTECTED] Examples]$ ls Controller SimpleApp [EMAIL PROTECTED] Examples]$ pwd /home/heron/Releases/gorm-1.1.0/Documentation/Examples [EMAIL PROTECTED] Examples]$ There are your simple examples. They've been there for a quite a while. Later, GJC And what if you dont know about Gorm or have decided not to use it? You might well want to try building a simple console app before a GUI one, or might want to build the GUI one by hand so as to get the feel of the code, rather than have it generated? I think the sample code should be included in the base development install, or perhaps with every dev tool? I also think the web site could do much better to highlight these resources. How about a developers intro page, with sample code actually on the web? I can find lots of other stuff on the site, but its the things that are essential for a first time user are not easy to find. And its the first time users who do not know either the software or the site that need these things. Experienced users have found the resources for themselves. This is the usual situation with open source projects - the info is all there, but you have to either be very lucky to stumble on it quickly, or spend months looking for a needle in a haystack. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
On 9/4/06, phil taylor [EMAIL PROTECTED] wrote: On Sun, 2006-09-03 at 21:08 -0700, Gregory John Casamento wrote: Phil, phil taylor [EMAIL PROTECTED]: If they cant bother to spend a couple of hours to knock up some sample programs, or build a few packages for the major distros (that arent four years out of date) what kind of message does that send to new users about the platform and the developers of it? I know the message it send to me - go somewhere else. In the Gorm release, since about 0.5.0 (most current release is 1.1) there are examples in the Documentation directory. They are covered in the Gorm manual which is also in that directory. [EMAIL PROTECTED] Examples]$ ls Controller SimpleApp [EMAIL PROTECTED] Examples]$ pwd /home/heron/Releases/gorm-1.1.0/Documentation/Examples [EMAIL PROTECTED] Examples]$ There are your simple examples. They've been there for a quite a while. Later, GJC And what if you dont know about Gorm or have decided not to use it? You might well want to try building a simple console app before a GUI one, or might want to build the GUI one by hand so as to get the feel of the code, rather than have it generated? I think the sample code should be included in the base development install, or perhaps with every dev tool? I also think the web site could do much better to highlight these resources. How about a developers intro page, with sample code actually on the web? I can find lots of other stuff on the site, but its the things that are essential for a first time user are not easy to find. And its the first time users who do not know either the software or the site that need these things. Experienced users have found the resources for themselves. This is the usual situation with open source projects - the info is all there, but you have to either be very lucky to stumble on it quickly, or spend months looking for a needle in a haystack. what about tests? they can be the most complete and up to date sample and documentation rolled into one. -- the thing i like with my linux pc is that i can sum up my complaints in 5 items ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 2006-08-30 19:44:27 +0200 Michael Thaler [EMAIL PROTECTED] wrote: On Wednesday 30 August 2006 14:41, [EMAIL PROTECTED] wrote: True, especially regarding Gecko. However, using libwww WOULD be an idea for writing a _very_ simple webbrowser. I was thinking about that, using libwww as a bundle, so it could be easily replaced in case WebKit (or something else that's better than libwww) came along. What is a _very_ simple webbrowser good for? I don't think many peope are interested in using such a browser if there are far better alternatives like Firefox and Konqueror. Put it this way: Why is there links, lynx or w3m, when there's Firefox or Opera? If it has a modular struture, it would be a starting point for a more complex web browser by replacing existing (simple) plugins with more sophisticated ones, like the rendering engine. The original browser could use libwww, if something like WebKit would become available, simply replace that plugin with one that uses WebKit. Additinal plugins, eg for bookmarks, history or a download managre can be added later on. Having the simple webbrowser grow into a full blown web browser suite. Additionally, as already pointed out, it can be embedded, eg. for context sensitive Help. It could also be used as a starting point for a CSS/HTML editor (think Nexus). -- Chris ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 2006-08-30 04:18:36 +0200 Jason Clouse [EMAIL PROTECTED] wrote: [...] I was working on a CoreFoundation implementation, but I hear that WebKit is totally abandoning Cocoa, so I'm not sure it makes a difference now. :-( Wouldn't make sense. They (that is, the WebKit developers) are said to drop Objective-C, moving to plain C++ instead, however. Doesn't make much sense either. Not to me at least. I think Gecko and libwww are poor choices, personally. True, especially regarding Gecko. However, using libwww WOULD be an idea for writing a _very_ simple webbrowser. I was thinking about that, using libwww as a bundle, so it could be easily replaced in case WebKit (or something else that's better than libwww) came along. -- Chris ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Chris wrote: Wouldn't make sense. They (that is, the WebKit developers) are said to drop Objective-C, moving to plain C++ instead, however. Doesn't make much sense either. Not to me at least. As far as I know, there are plans to merge KHTML and WebCore into one project (named Unity or something like that) to join forces. See e.g.: http://lists.kde.org/?l=kfm-develm=115274146924256w=2 In that context it would make sense to remove Cocoa dependencies. Then WebKit was just a Obj-C wrapper around the C++ only WebCore/KHTML stuff. Axel ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Nikolaus Waxweiler [EMAIL PROTECTED] Yes, I noticed the global page protection already and am still waiting for write-access. Other methods for fighting spam bots, like captcha codes on registration (or maybe on every edit), should lower the entry bar enough to get more people to contribute. CAPTCHAs would lock more people out: http://www.w3.org/TR/turingtest/ And please, strange markup is no argument, Mediawiki's markup is quite simple and there are quick references if you need them. Mediawiki's markup is not html (known by web authors), not autogsdoc or texinfo (used for current GS docs) and not wiki TextFormattingRules. Having to keep open a quick-reference guide is a poor substitute. The web API documentation is currently synchronised. No, not really. Just look at http://mediawiki.gnustep.org/index.php/AppKit and http://mediawiki.gnustep.org/index.php/Foundation. They consist mostly of red links. By 'The web API documentation', I meant the copies on www.gnustep.org - I have no access to mediawiki.gnustep.org and think it requires human copy-paste to update. It seems like it could be fairly easy make the main site searchable without search engines, if that's something which people want. It's just not been a terribly high priority. Maybe, but it would be easier to just have everything in the same place. So let's kill mediawiki and the appdb and start them again as parts of www. [...] Then, there aren't enough easy-to-use practical guides to creating new applications for either gnustep users getting started with development, or non-obJC developers from other systems (Perl, Python, Ruby, Lisps..., or the distribution packagers) looking to help GNUstep. There are several introductory guides and even a few videos iirc. If newbies or developers want a more in-depth look at GNUstep, they can get Cocoa books. Yes, there are some guides and videos, but not *enough* for new user-developers and particularly not enough for other system developers. Referring them to Cocoa books seems unhelpful: GNUstep doesn't implement it all and some things are done different. -- MJR/slef ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
In article [EMAIL PROTECTED], Helge Hess [EMAIL PROTECTED] wrote: On Aug 27, 2006, at 17:51, Doc O'Leary wrote: I agree with your premise, but not the conclusion. Yes, the Linux market is tiny, but as a developer I would gladly deploy there if the effort were also as tiny to port my Mac software. Hm, ok. Why would you want to do that? Hm, ok. Why *wouldn't* you want to do that? I want people to use my software, and I don't really care if they're using a Mac. If it were trivial to make software available to Linux users, then I don't see how it benefits anyone other than Apple to keep it off other platforms. As a proprietary software developer, why would I port to a system which isn't used? It doesn't matter how easy it is. Lets say porting Delicious Library to GNUstep/Linux would take 30 hours which would be a very tiny effort. Now I don't even agree with your premise. 30 hours is *not* a tiny effort. A tiny effort is ticking a check box like NeXT allowed. As I stated, if there were a way I could cross-compile and/or have my linked application just work with GNUstep, it would make a huge difference in how much software is available. As it is, there isn't even a lot of interest in getting existing open source Mac/Cocoa software running on GNUstep. PS: if you would make it possible to port such Cocoa applications in less than a week to GNOME or KDE, it would certainly make sense for small scale developers. So feel free to add this to my list :-): c) reasonably easy and convenient KDE/GNOME porting for Cocoa developers I will definitely agree that GNUstep could do wonders as a bridging technology. Like many, I was sorely disappointed with the loss of Yellow Box, and I have already stated that a better focus for GNUstep would be for portability rather than as a primary platform. So while Linux might not be that attractive a market financially, technically it makes a good target. Hm, then you didn't get my initial/basic point :-) The former is the driving incentive for most Cocoa developers (as mentioned, very little OpenSource Cocoa apps, plenty of shareware style ones). If you want to get them, you need to make it attractive financially. Then my point was lost, too! I'm not saying Linux is a financially rewarding target market, but rather that it is technically a good first target for portability. Once the direction is set and that initial baby step is taken, portability can definitely expand to Windows, or any other platform that *is* financially viable. -- My personal UDP list: 127.0.0.1, 4ax.com, buzzardnews.com, googlegroups.com, heapnode.com, localhost, x-privat.org ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 8/30/06, Axel 'Mikesch' Katerbau [EMAIL PROTECTED] wrote: [...] In that context it would make sense to remove Cocoa dependencies. Then WebKit was just a Obj-C wrapper around the C++ only WebCore/KHTML stuff. No, it does not (at least to me) because THAT is where WebKit used to come FROM. WebKit USED to be 'just an ObjC wrapper around KHTML's C++ code. I do not get it, why they would want to go back there, because a) that's where it came from b) KHTML is still available separately c) WebKit's first and foremost platform is OSX But maybe I'm just missing the point. Wouldn't be the first time :-) -- Chris ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
CAPTCHAs would lock more people out: http://www.w3.org/TR/turingtest/ A captcha is just one solution of many. The URL you referenced contains alternate methods like logic puzzles. Mediawiki's markup is not html (known by web authors), not autogsdoc or texinfo (used for current GS docs) and not wiki TextFormattingRules. Having to keep open a quick-reference guide is a poor substitute. You can use HTML, plus the most commonly used markup is simple to remember. I looked at http://www.usemod.com/cgi-bin/wiki.pl?TextFormattingRules and the markup seems to be similar. By 'The web API documentation', I meant the copies on www.gnustep.org - I have no access to mediawiki.gnustep.org and think it requires human copy-paste to update. Yes, I was thinking about a script that automatically updates the wiki API documentation, but.. yeah, it might be better to use the wiki as a place for general information about classes, code snippets, tricks, etc. So let's kill mediawiki and the appdb and start them again as parts of www. You're proposing to leave everything as it currently is :P Yes, there are some guides and videos, but not *enough* for new user-developers and particularly not enough for other system developers. Referring them to Cocoa books seems unhelpful: GNUstep doesn't implement it all and some things are done different. GNUstep and Cocoa share the same ideas and concepts. Unless you use Mac OS X specific things such as CoreAnimation, the differences are negligible or at least manageable. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On Wednesday 30 August 2006 14:41, [EMAIL PROTECTED] wrote: True, especially regarding Gecko. However, using libwww WOULD be an idea for writing a _very_ simple webbrowser. I was thinking about that, using libwww as a bundle, so it could be easily replaced in case WebKit (or something else that's better than libwww) came along. What is a _very_ simple webbrowser good for? I don't think many peope are interested in using such a browser if there are far better alternatives like Firefox and Konqueror. In case you don't know, there is a free webbrowser based on WebKit and written in Cocoa: Shiira (http://hmdt-web.net/shiira/en). It's BSD licensed. I cannot try it because I don't have a Mac. But a browser like this (or Safari) would make gnustep much more interesting for users (and developers are also users and often help with projects that they use). Personally, I kept a eye on gnustep for a long time and and every now and then I think about writing a small gnustep app (I would like to write a simple painting app like KolourPaint). But I always abolished the idea because I cannot use gnustep as my desktop (and gnustep apps are unsuable under KDE) and developed for KDE instead. But I still hope that gnustep is useable as a desktop some day (and for me that means mainly, that basic applications that I use every day like an email programm and a browser are available). At this point I probably will start to develop for gnustep. Greetings, Michael ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On Aug 30, 2006, at 1:44 PM, Michael Thaler wrote: On Wednesday 30 August 2006 14:41, [EMAIL PROTECTED] wrote: True, especially regarding Gecko. However, using libwww WOULD be an idea for writing a _very_ simple webbrowser. I was thinking about that, using libwww as a bundle, so it could be easily replaced in case WebKit (or something else that's better than libwww) came along. What is a _very_ simple webbrowser good for? I don't think many peope are interested in using such a browser if there are far better alternatives like Firefox and Konqueror. It could be useful in embedded contexts -- e.g., a class documentation viewer for ProjectCenter, HTML email viewing in GNUMail, application help, etc.. If libwww is something relatively lean and mean, so much the better. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
[...] In that context it would make sense to remove Cocoa dependencies. Then WebKit was just a Obj-C wrapper around the C++ only WebCore/KHTML stuff.No, it does not (at least to me) because THAT is where WebKit used to come FROM. WebKit USED to be 'just an ObjC wrapper around KHTML's C++ code.I do not get it, why they would want to go back there, becausea) that's where it came fromb) KHTML is still available separatelyc) WebKit's first and foremost platform is OSX But maybe I'm just missing the point. Wouldn't be the first time :-)--ChrisActually, since the full WebKit was released open source, there has been a community-driven movement to make it platform agnostic. Much of the code has been refactored so that the core _javascript_ and web stuff is done in platform independent C++, and then there are different supported platforms for the presentation and rendering including Cocoa, win32, gtk+, and the Nokia's S60 phone operating system. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 2006-08-28 18:43:57 +0200 Nikolaus Waxweiler [EMAIL PROTECTED] wrote: [...] The web API documentation is currently synchronised. No, not really. Just look at http://mediawiki.gnustep.org/index.php/AppKit and http://mediawiki.gnustep.org/index.php/Foundation. They consist mostly of red links. [...] That was my doing and as much as I hate to admit it, I do have a day job that keeps me from updating the page(s) more frequently :-( -- Chris ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 8/29/06, Chris Vetter [EMAIL PROTECTED] wrote: On 2006-08-28 19:57:06 +0200 Riccardo [EMAIL PROTECTED] wrote: [...] incomplete libraries? That's open to debate. If you compare it to Cocoa, yes, it's incomplete. Question is how far the completeness is supposed to go (with respect to Cocoa). [...] If we just implement the OpenStep specification (and ignore cocoa, which added nice thing and ruined others) we have enough power to write our web browser (yes, seriously, it is time and resources which lack, not frameworks), an office suite, an imaging application, a 3d application it was done on OpenStep and we even have stuff which was not in OS. [...] I beg to differ. Which framework(s) are you referring to regarding a web browser? Yes, there basically are 3 to pick from (libwww, Mozilla's engine and WebKit) but the first two would need wrapping and the latter is a PITA to port (I tried several times and got stuck due to references to Apple specific frameworks) plus the latter two are (currently) using GTK/GDK, which -- for me -- is a 'no go.' i looked at amaya and i got stuck in the wxwindows parts too. but it looks more promising than the others. maybe we need to come up with an alternative engine. its ironic considering the www was invented on a next cube. -- things i hate about my linux pc: 1. it takes more than a second to boot up 2. keeps asking about filenames and directories 3. does not remember what i was working on yesterday 4. does not remember all the changes i have ever made 5.cannot figure out necessary settings by itself ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 2006-08-29 09:40:22 +0200 Rogelio Serrano [EMAIL PROTECTED] wrote: [...] i looked at amaya and i got stuck in the wxwindows parts too. but it looks more promising than the others. maybe we need to come up with an alternative engine. its ironic considering the www was invented on a next cube. Yes, and Nexus (aka WorldWideWeb.app) was using what nowadays is called libwww. -- Chris ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 8/29/06, Chris Vetter [EMAIL PROTECTED] wrote: On 2006-08-29 09:40:22 +0200 Rogelio Serrano [EMAIL PROTECTED] wrote: [...] i looked at amaya and i got stuck in the wxwindows parts too. but it looks more promising than the others. maybe we need to come up with an alternative engine. its ironic considering the www was invented on a next cube. Yes, and Nexus (aka WorldWideWeb.app) was using what nowadays is called libwww. why dont we try it then? -- things i hate about my linux pc: 1. it takes more than a second to boot up 2. keeps asking about filenames and directories 3. does not remember what i was working on yesterday 4. does not remember all the changes i have ever made 5.cannot figure out necessary settings by itself ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 2006-08-29 03:40:22 -0400 Rogelio Serrano [EMAIL PROTECTED] wrote: i looked at amaya and i got stuck in the wxwindows parts too. but it looks more promising than the others. I don't kmow about that. The rendering in Amaya is terrible and many pages just don't work. Charles -- DPRINTK(Last time you were disconnected, how about now?\n); linux-2.6.6/drivers/net/tokenring/ibmtr.c RFC3156 defines security multipart formats for MIME with OpenPGP. pgpZQJlVhBCGJ.pgp Description: PGP signature ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: GNUstep paste board (Was: really attracting developers)
On 2006-08-28 15:59:32 -0400 Fred Kiefer [EMAIL PROTECTED] wrote: I am a bit unsure, how much of this is already a standard or at least widely used. But having image exchange in GNUstep would surely be nice. I know at least it is implemented in GTK and QT because image exchange between GTK and QT apps works. One problem here is that GNUstep can only convert NSImages to TIFF, but not to PNG. Perhaps it would be a start to be able to read PNG from the pasteboard, which should not be too hard to implement. I wonder if just by copy TIFF data to the X clipboard is good enough. As for the font format, I think it is actually a very bad idea to copy the real font, in what ever format, onto the clipboard. What should be put there are the font properties, just like GNUstep does it :-) Agreed. I was just paraphrasing the spec. DND will be another big step. Here a whole rewrite is needed. When we tried to integrate GNUstep DND with X we did get it completely wrong the first time. A working DND system will certainly attract more less technical users. Yes, this is the biggest remaining problem, but it does not look to hard to resolve, somebody will just have to take a deep look into a PS specification and then hard code it into GSStreamContext. I see. Charles -- printk (Kicking board.\n); linux-2.6.6/drivers/net/lp486e.c RFC3156 defines security multipart formats for MIME with OpenPGP. pgpfphhIHdx5t.pgp Description: PGP signature ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: GNUstep paste board (Was: really attracting developers)
Great! Thank you for the code I have a closer look soon.
Fred
Matt Rice schrieb:
--- Fred Kiefer [EMAIL PROTECTED] wrote:
DND will be another big step. Here a whole rewrite
is needed. When we
tried to integrate GNUstep DND with X we did get it
completely wrong the
first time.
a while back i tried getting gnustep to work as a x
'dragging source' as we call it.. it still uses the
outdated xdnd code but attached is what i ended up
with and a dragging destination which also uses the
outdated xdnd code from gnustep..
the whole thing inside the case SelectionRequest:
needs to be rewritten doesnt do any of the conversion
it should (file uris should have file:// etc and other
conversion methods for other stuff...)
i recall there being alot more to the patch than this
but this is all there seemed to be... so hopefully its
all here...
i also remember there being times when the code just
magically didnt work..
hope it helps, dont forsee finishing it anytime soon
if someone wants to pick it up...
__
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
Index: Source/x11/XGServerEvent.m
===
--- Source/x11/XGServerEvent.m(revision 22775)
+++ Source/x11/XGServerEvent.m(working copy)
@@ -30,6 +30,7 @@
#include AppKit/NSGraphics.h
#include AppKit/NSMenu.h
#include AppKit/NSWindow.h
+#include AppKit/NSPasteboard.h
#include Foundation/NSException.h
#include Foundation/NSArray.h
#include Foundation/NSDictionary.h
@@ -1294,10 +1295,34 @@
break;
case SelectionRequest:
- NSDebugLLog(@NSEvent, @%d SelectionRequest\n,
- xEvent.xselectionrequest.requestor);
+ {
+ DndClass dnd = xdnd();
+ /* ugly hack alert delete all this stuff please*/
+ NSPasteboard *pb = [NSPasteboard pasteboardWithName:NSDragPboard];
+ NSArray *types = [pb types];
+ NSLog(@%@, [pb types]);
+ NSData *data;
+ if ([types count])
+ {
+ id uh = [pb propertyListForType:[types objectAtIndex:0]];
+ if ([uh isKindOfClass:[NSArray class]])
+ {
+ if ([uh count])
+ {
+ uh = [uh objectAtIndex:0];
+
+ }
+ NSLog(@%@, uh);
+ data = [uh dataUsingEncoding:NSASCIIStringEncoding
allowLossyConversion:YES];
+ }
+ xdnd_selection_send(dnd, xEvent.xselectionrequest, (unsigned
char *)[data bytes], [data length]);
+ }
+
+ NSDebugLLog(@NSEvent, @%d SelectionRequest\n,
+ xEvent.xselectionrequest.requestor);
+ }
break;
-
+
// We shouldn't get here unless we forgot to trap an event above
default:
#ifdef XSHM
Index: Source/x11/XGDragView.m
===
--- Source/x11/XGDragView.m (revision 22775)
+++ Source/x11/XGDragView.m (working copy)
@@ -276,6 +276,7 @@
break;
case GSAppKitDraggingEnter:
+ xdnd_set_selection_owner(dnd, dragWindev-ident, typelist[0]);
xdnd_send_enter(dnd, dWindowNumber, dragWindev-ident, typelist);
xdnd_send_position(dnd, dWindowNumber, dragWindev-ident,
GSActionForDragOperation (dragMask operationMask),
___
Discuss-gnustep mailing list
Discuss-gnustep@gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Hey, On Tuesday, August 29, 2006, at 09:29 AM, Chris Vetter wrote: [...] If we just implement the OpenStep specification (and ignore cocoa, which added nice thing and ruined others) we have enough power to write our web browser (yes, seriously, it is time and resources which lack, not frameworks), an office suite, an imaging application, a 3d application it was done on OpenStep and we even have stuff which was not in OS. [...] I beg to differ. Which framework(s) are you referring to regarding a web browser? Yes, there basically are 3 to pick from (libwww, Mozilla's engine and WebKit) but the first two would need wrapping and the latter is a PITA to port (I tried several times and got stuck due to references to Apple specific frameworks) plus the latter two are (currently) using GTK/GDK, which -- for me -- is a 'no go.' Actually what I meant is that gnustep-core is complete enough that you could write a webbrowser from scratch. Despite what people think, it can be done (iCab is a browser for macintosh written by two brothers, very heavy-weight, it run on 68k for a long time, it supports javascript, jscript and lately CSS is in the works too and mind, it is done in classic mac and nowadays carbon, thus far more primitive stuff). There is also a browser on windows wirten from scratch wich is reasonably nice. If you want to stir up the browser argument again, I think web-core is a bit of a dead-end. We could hack on libwww (many browser did that), or hack on the links graphics core, which even supports some limited javascript and is in pure C. Or try to wrap gecko like it was done with camino. I think that if we make our own wrappers it could be a nice result and possibly even without that monster called obj-c++. Cheers, Riccardo ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 2006-08-29 17:07:47 -0400 Michael Hanni [EMAIL PROTECTED] wrote: Just throwing an idea out here: would it be possible to bind directly to KHTML and skip the WebKit framework nightmare? [1] Somebody did this a few years ago, I think. But it seems a shame not to ride on Apple's coattails with WebKit. After all, they have the manpower for QA and development, and plenty of web developers check that their stuff works in Safari. I was working on a CoreFoundation implementation, but I hear that WebKit is totally abandoning Cocoa, so I'm not sure it makes a difference now. :-( I think Gecko and libwww are poor choices, personally. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 8/30/06, Jason Clouse [EMAIL PROTECTED] wrote: On 2006-08-29 17:07:47 -0400 Michael Hanni [EMAIL PROTECTED] wrote: Just throwing an idea out here: would it be possible to bind directly to KHTML and skip the WebKit framework nightmare? [1] Somebody did this a few years ago, I think. But it seems a shame not to ride on Apple's coattails with WebKit. After all, they have the manpower for QA and development, and plenty of web developers check that their stuff works in Safari. I was working on a CoreFoundation implementation, but I hear that WebKit is totally abandoning Cocoa, so I'm not sure it makes a difference now. :-( I think Gecko and libwww are poor choices, personally. poor choices but the only available ones. and they can be improved. Maybe we should try doing both. Like what i was told in an extreme programming seminar, go ahead and fail. i would like to try libwww again. but then the complexity of it all makes me look for a simpler alternative not just to the browser but the whole www. -- things i hate about my linux pc: 1. it takes more than a second to boot up 2. keeps asking about filenames and directories 3. does not remember what i was working on yesterday 4. does not remember all the changes i have ever made 5.cannot figure out necessary settings by itself ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 8/30/06, Rogelio Serrano [EMAIL PROTECTED] wrote: On 8/30/06, Jason Clouse [EMAIL PROTECTED] wrote: On 2006-08-29 17:07:47 -0400 Michael Hanni [EMAIL PROTECTED] wrote: Just throwing an idea out here: would it be possible to bind directly to KHTML and skip the WebKit framework nightmare? [1] Somebody did this a few years ago, I think. But it seems a shame not to ride on Apple's coattails with WebKit. After all, they have the manpower for QA and development, and plenty of web developers check that their stuff works in Safari. I was working on a CoreFoundation implementation, but I hear that WebKit is totally abandoning Cocoa, so I'm not sure it makes a difference now. :-( I think Gecko and libwww are poor choices, personally. poor choices but the only available ones. and they can be improved. Maybe we should try doing both. Like what i was told in an extreme programming seminar, go ahead and fail. i would like to try libwww again. but then the complexity of it all makes me look for a simpler alternative not just to the browser but the whole www. i was wondering too if the web interface and the desktop interface was merged by design. two birds with one stone. -- things i hate about my linux pc: 1. it takes more than a second to boot up 2. keeps asking about filenames and directories 3. does not remember what i was working on yesterday 4. does not remember all the changes i have ever made 5.cannot figure out necessary settings by itself ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: GNUstep paste board (Was: really attracting developers)
On 8/28/06, Charles Philip Chan [EMAIL PROTECTED] wrote: [...] It works for basic text, but not for other data types. I can't speak for other platforms, but on an X based system I think this should happen: (1) If a bitmap graphics is cut or copied, it should make a png copy to the X clipboard. (2) If it is a font, copy it to the X clipboard as a TTf. [...] Other data types should follow as the standard evolves. Actually, you CAN copy images using the pasteboard. I frequently use that functionality. Don't know about fonts, but I'm pretty sure it would be possible. -- Chris ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: GNUstep paste board (Was: really attracting developers)
On 2006-08-28 03:38:43 -0400 Chris B. Vetter [EMAIL PROTECTED] wrote: Actually, you CAN copy images using the pasteboard. I frequently use that functionality. Don't know about fonts, but I'm pretty sure it would be possible. Only between GNUstep apps. I just try a couple of experiments: (1) Preview.app- OpenOffice. Results: Requested Clipboard Format not Available. (2) GWorkspace- OpenOffice. Results: Nothing. Charles -- printk (%s: This looks like a LART board to me.\n,module_name); linux-2.6.6/drivers/mtd/devices/lart.c RFC3156 defines security multipart formats for MIME with OpenPGP. pgpKrzk8MAggb.pgp Description: PGP signature ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
In article [EMAIL PROTECTED], Rogelio Serrano [EMAIL PROTECTED] wrote: Whats keeping other developers from gnustep? Depends on who those other developers are. I can only speak as someone who 10 years ago used GNUstep (such as it was) on Linux, but was drawn by commercial interests to focus on NeXT/Apple development. For me, it's a question of what would get me *back* into GNUstep development. I'm the essentially the low bar; if you can't convince me then you're in trouble. incomplete ide? More useful to discuss is how to take an OpenStep project developed on Mac OS X and get it working on a GNUstep system. I don't at all poo-poo the work on ProjectCenter, but if you really address where new development is coming from you'll probably find that a lot of people are cutting their ObjC teeth on Xcode. More useful to me would be having Xcode cross-compile and/or link to produce a binary that runs on a GNUstep system. incomplete nextstep based system? incomplete libraries? Not in any major way that a developer couldn't solve. What matters more is the lack of a market, because I'm not going to expend 10x the effort to develop for a user base that is 100x smaller than even the Mac's. What interests me still about GNUstep is the possibility of taking my Mac apps to a place where Apple won't go, say like the Nokia 770. Given the dramatic price drop that has occurred since NeXT took over Apple (yay, no more $5000 developer tools!), I think GNUstep is currently best served by shifting focus slightly to be a deployment platform more than a development platform. -- My personal UDP list: 127.0.0.1, 4ax.com, buzzardnews.com, googlegroups.com, heapnode.com, localhost, x-privat.org ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On Mon, 28 Aug 2006 14:16:30 +0200, MJ Ray [EMAIL PROTECTED] wrote: You've answered your own question. The wiki is not good enough yet, underattended and stumpy. Erm, moving as much stuff as possible to the wiki was my answer to the underattended and stumpy issue. Etoile and other projects like the Beep Media Player also use a wiki as the main homepage. No, they can't. Mediawiki has some login system that locks people out and a strange markup which discourages old c2.com WikiFarmers like me. Yes, I noticed the global page protection already and am still waiting for write-access. Other methods for fighting spam bots, like captcha codes on registration (or maybe on every edit), should lower the entry bar enough to get more people to contribute. And please, strange markup is no argument, Mediawiki's markup is quite simple and there are quick references if you need them. The web API documentation is currently synchronised. No, not really. Just look at http://mediawiki.gnustep.org/index.php/AppKit and http://mediawiki.gnustep.org/index.php/Foundation. They consist mostly of red links. It seems like it could be fairly easy make the main site searchable without search engines, if that's something which people want. It's just not been a terribly high priority. Maybe, but it would be easier to just have everything in the same place. 3. other sites seem to break their inbound links so often that it's hard to keep a working set of links on www.gnustep.org Just pull the interesting stuff into the wiki :P. Then, there aren't enough easy-to-use practical guides to creating new applications for either gnustep users getting started with development, or non-obJC developers from other systems (Perl, Python, Ruby, Lisps..., or the distribution packagers) looking to help GNUstep. There are several introductory guides and even a few videos iirc. If newbies or developers want a more in-depth look at GNUstep, they can get Cocoa books. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Hey, On Friday, August 25, 2006, at 01:57 PM, Chris Vetter wrote: On 2006-08-25 12:05:48 +0200 Rogelio Serrano [EMAIL PROTECTED] wrote: Whats keeping other developers from gnustep? That's the crucial question, isn't it? I think we had this thread before.. . and even before before incomplete ide? Gorm, while probably not finished, is very stable and usable. While Gorm is not an IDE.. it is very nice and works well. I have to add that ProjectCenter, while being SERIOUSLY incomplete in some parts, works well in what it delivers, so for many project it is just fine and it is reasonably stable. SO I don'tb uy this argument. We may make progress... but we are better off than other toolkits even. incomplete libraries? That's open to debate. If you compare it to Cocoa, yes, it's incomplete. Question is how far the completeness is supposed to go (with respect to Cocoa). I think the libraries are pretty complete. Try to code something (not to port some crap) and check if you can't do that because of a method not implemented. If it is in the OpenStep, then we should implement it, otherwise just start coding. If we just implement the OpenStep specification (and ignore cocoa, which added nice thing and ruined others) we have enough power to write our web browser (yes, seriously, it is time and resources which lack, not frameworks), an office suite, an imaging application, a 3d application it was done on OpenStep and we even have stuff which was not in OS. i don't buy the general applications unavailability argument. we are talking about people who want to create apps under gnustep. There aren't many usable GNUstep related applications. Not from a user's point of view. Most notably a web browser is needed. I agree. Several desktop applications are missing or incomplete. There are bugs here and there. But it is not the gnustep core that is holding back applications. -Riccardo ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
hmmm.. On Friday, August 25, 2006, at 07:12 PM, Chris Vetter wrote: without more ado then: http://nextbuntu.wordpress.com/ Looks to me like this guy is divorced from reality. do you know who that guys? He needs medical support. -R ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
On Aug 27, 2006, at 17:51, Doc O'Leary wrote: Summary: having Cocoa compatibility (which is getting harder every day as MacOS advances, just think ObjC 2.0) for Linux is _not_ a selling point. The majority of Cocoa developers simply don't want to deploy their desktop applications to Linux/BSD. I agree with your premise, but not the conclusion. Yes, the Linux market is tiny, but as a developer I would gladly deploy there if the effort were also as tiny to port my Mac software. Hm, ok. Why would you want to do that? It currently is not, and that is what makes GNUstep not as attractive as it could be. As a proprietary software developer, why would I port to a system which isn't used? It doesn't matter how easy it is. Lets say porting Delicious Library to GNUstep/Linux would take 30 hours which would be a very tiny effort. Let assume that a developer hour is just $100, this would make just $3000 for a port. Not much at all if the software makes a few hundred thousands of bucks. You would need to sell to just 75 people to break even ($3000 / $40). You'll have a hard time finding 75 people using GNUstep on a desktop. And it will be even more difficult to find 5 people actually willing to _pay_ for the application. Now of course it could be done by a few Cocoa devs as a marketing gag or just because they can. But still it wouldn't be any significant motivation. PS: if you would make it possible to port such Cocoa applications in less than a week to GNOME or KDE, it would certainly make sense for small scale developers. So feel free to add this to my list :-): c) reasonably easy and convenient KDE/GNOME porting for Cocoa developers IMHO there are two spaces which can be explored if you want to advance the GNUstep community: a) reasonably easy and convenient Windows porting for Cocoa developers b) server stuff I completely agree here. I will note, however, that in getting to Windows it *should* be an easier first step to run Cocoa/Mac apps on another Unix system. I don't think its more difficult in itself (possibly porting Aqua effects stuff to Windows graphics stuff is actually easier due to better graphics support, don't know). But I do think its more difficult for _Unix_ developers (all current GNUstep devs ;-). But then there are plenty of Windows devs. Should be easy to acquire them to reproduce MacOS on Windows, no? ;-) So while Linux might not be that attractive a market financially, technically it makes a good target. Hm, then you didn't get my initial/basic point :-) The former is the driving incentive for most Cocoa developers (as mentioned, very little OpenSource Cocoa apps, plenty of shareware style ones). If you want to get them, you need to make it attractive financially. Greets, Helge -- Helge Hess http://docs.opengroupware.org/Members/helge/ ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: GNUstep paste board (Was: really attracting developers)
Charles Philip Chan schrieb: On 2006-08-27 15:41:37 -0400 Fred Kiefer [EMAIL PROTECTED] wrote: Could you please detail the problem you found with GNUstep paste board? At least in a Unix/Linux environment I would expect GNUstep to play nicely with other apps that follow the basic rules. If you found a problem I am willing to look into this. It works for basic text, but not for other data types. I can't speak for other platforms, but on an X based system I think this should happen: (1) If a bitmap graphics is cut or copied, it should make a png copy to the X clipboard. (2) If it is a font, copy it to the X clipboard as a TTf. This is in accordance with the Free Desktop standard. Please refer to this page: http://www.freedesktop.org/wiki/Standards_2fClipboardsWiki Other data types should follow as the standard evolves. I am a bit unsure, how much of this is already a standard or at least widely used. But having image exchange in GNUstep would surely be nice. One problem here is that GNUstep can only convert NSImages to TIFF, but not to PNG. Perhaps it would be a start to be able to read PNG from the pasteboard, which should not be too hard to implement. As for the font format, I think it is actually a very bad idea to copy the real font, in what ever format, onto the clipboard. What should be put there are the font properties, just like GNUstep does it :-) Yes, here there is still much to do. The strange thing is that up to now people only rarely asked for this. Strange indeed since DND is used widely in NeXTStep. :-) DND will be another big step. Here a whole rewrite is needed. When we tried to integrate GNUstep DND with X we did get it completely wrong the first time. It almost is there :-) Really, but we need to improve on it. I think the problem stems from incorrect Postscript font names in the generated ps data. I have tested it out and Ghostscript chokes on them. Yes, this is the biggest remaining problem, but it does not look to hard to resolve, somebody will just have to take a deep look into a PS specification and then hard code it into GSStreamContext. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: GNUstep paste board (Was: really attracting developers)
--- Fred Kiefer [EMAIL PROTECTED] wrote:
DND will be another big step. Here a whole rewrite
is needed. When we
tried to integrate GNUstep DND with X we did get it
completely wrong the
first time.
a while back i tried getting gnustep to work as a x
'dragging source' as we call it.. it still uses the
outdated xdnd code but attached is what i ended up
with and a dragging destination which also uses the
outdated xdnd code from gnustep..
the whole thing inside the case SelectionRequest:
needs to be rewritten doesnt do any of the conversion
it should (file uris should have file:// etc and other
conversion methods for other stuff...)
i recall there being alot more to the patch than this
but this is all there seemed to be... so hopefully its
all here...
i also remember there being times when the code just
magically didnt work..
hope it helps, dont forsee finishing it anytime soon
if someone wants to pick it up...
__
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
xdnd-dest.tar.gz
Description: 296406374-xdnd-dest.tar.gz
Index: Source/x11/XGServerEvent.m
===
--- Source/x11/XGServerEvent.m (revision 22775)
+++ Source/x11/XGServerEvent.m (working copy)
@@ -30,6 +30,7 @@
#include AppKit/NSGraphics.h
#include AppKit/NSMenu.h
#include AppKit/NSWindow.h
+#include AppKit/NSPasteboard.h
#include Foundation/NSException.h
#include Foundation/NSArray.h
#include Foundation/NSDictionary.h
@@ -1294,10 +1295,34 @@
break;
case SelectionRequest:
- NSDebugLLog(@NSEvent, @%d SelectionRequest\n,
- xEvent.xselectionrequest.requestor);
+ {
+ DndClass dnd = xdnd();
+ /* ugly hack alert delete all this stuff please*/
+ NSPasteboard *pb = [NSPasteboard pasteboardWithName:NSDragPboard];
+ NSArray *types = [pb types];
+ NSLog(@%@, [pb types]);
+ NSData *data;
+ if ([types count])
+ {
+ id uh = [pb propertyListForType:[types objectAtIndex:0]];
+ if ([uh isKindOfClass:[NSArray class]])
+ {
+ if ([uh count])
+ {
+ uh = [uh objectAtIndex:0];
+
+ }
+ NSLog(@%@, uh);
+ data = [uh dataUsingEncoding:NSASCIIStringEncoding
allowLossyConversion:YES];
+ }
+ xdnd_selection_send(dnd, xEvent.xselectionrequest, (unsigned
char *)[data bytes], [data length]);
+ }
+
+ NSDebugLLog(@NSEvent, @%d SelectionRequest\n,
+ xEvent.xselectionrequest.requestor);
+ }
break;
-
+
// We shouldn't get here unless we forgot to trap an event above
default:
#ifdef XSHM
Index: Source/x11/XGDragView.m
===
--- Source/x11/XGDragView.m (revision 22775)
+++ Source/x11/XGDragView.m (working copy)
@@ -276,6 +276,7 @@
break;
case GSAppKitDraggingEnter:
+ xdnd_set_selection_owner(dnd, dragWindev-ident, typelist[0]);
xdnd_send_enter(dnd, dWindowNumber, dragWindev-ident, typelist);
xdnd_send_position(dnd, dWindowNumber, dragWindev-ident,
GSActionForDragOperation (dragMask operationMask),
___
Discuss-gnustep mailing list
Discuss-gnustep@gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: Re: GNUstep paste board (Was: really attracting developers)
On 8/28/06, Charles Philip Chan [EMAIL PROTECTED] wrote: On 2006-08-27 15:41:37 -0400 Fred Kiefer [EMAIL PROTECTED] wrote: Yes, here there is still much to do. The strange thing is that up to now people only rarely asked for this. Strange indeed since DND is used widely in NeXTStep. :-) Well, to be fair, drag and drop *between gnustep apps* works perfectly; and the GNUstep/OpenStep pasteboard system is way ahead the X dnd stuff. What doesn't work well is drag and drop (well, afaik only text is supported) between a gnustep app and another X application.. -- Nicolas Roard I love deadlines. I like the whooshing sound they make as they fly by. -- Douglas Adams ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
GNUstep needs to be both. --Gregory John Casamento- Original Message From: Philippe C.D. Robert [EMAIL PROTECTED]To: Rogelio Serrano [EMAIL PROTECTED]Cc: GNUstep Discussion discuss-gnustep@gnu.orgSent: Saturday, August 26, 2006 3:47:02 PMSubject: Re: really attracting developersHi,On 25.08.2006, at 12:05, Rogelio Serrano wrote: Whats keeping other developers from gnustep? incomplete ide? incomplete nextstep based system? incomplete libraries? i don't buy the general applications unavailability argument. we are talking about people who want to create apps under gnustep. i don't buy the appearance argument either.It's funny (and sad...) that this discussion pops up every one andthen since years ...IMO it is because some GNUstep developers don't want to see GNUstepbeing a complete desktop solution (for X11 based systems), insteadthey favour the cross-platform API path, whereas others don't want towrite major applications for GNUstep because it is not a completedesktop and thus has no target audience - sadly GNUstep apps justdon't integrate with other desktops. It is kind of a chicken and eggsituation (yes I am aware of Etoile, but IMO I doubt this projectwill have a big impact on GNUstep in terms of attracting new users ordevelopers - unless it gets integrated into GNUstep itself).This issue has never been resolved and as a consequence the projectdoes not really make progress in terms of attracting new users ordevelopers. Worse, while Java, .NET(Mono) or even gtk and Qt madeprogress in terms of installed base, functionality and maturity,GNUstep simply did not. Even more worse, you cannot use GNUstep to deploy OS X code on other OSes since the beasts are too different,e.g. there are many more APIs on OSX which are being used other thanCocoa when developing OS X apps. Not even the Objective-C runtime is100% compatible (which will become a real issue with Leopard, I amafraid).So tell me one reason why I should write code for/with GNUstep unlessI am an Objective-C aficionado .. (which I am, actually :). IMOGNUstep needs to change its focus otherwise it will always remain oneof those many nice toolkits which will - unfortunately - never makeit to the real surface...just my $0.02... and now hit me ;-)-Phil--Philippe C.D. Roberthttp://www.nice.ch/~phip___Discuss-gnustep mailing listDiscuss-gnustep@gnu.orghttp://lists.gnu.org/mailman/listinfo/discuss-gnustep___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
GNUstep paste board (Was: really attracting developers)
Hi Charles, Charles Philip Chan schrieb: I agree with this. The show stoppers that I found with GNUstep are: (1) A paste board that play nicely with the native clipboard system. Could you please detail the problem you found with GNUstep paste board? At least in a Unix/Linux environment I would expect GNUstep to play nicely with other apps that follow the basic rules. If you found a problem I am willing to look into this. (2) Drap and drop that works with the native system. Yes, here there is still much to do. The strange thing is that up to now people only rarely asked for this. (3) A print system that works. It almost is there :-) Really, but we need to improve on it. Fred ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: GNUstep paste board (Was: really attracting developers)
On 2006-08-27 15:41:37 -0400 Fred Kiefer [EMAIL PROTECTED] wrote: Could you please detail the problem you found with GNUstep paste board? At least in a Unix/Linux environment I would expect GNUstep to play nicely with other apps that follow the basic rules. If you found a problem I am willing to look into this. It works for basic text, but not for other data types. I can't speak for other platforms, but on an X based system I think this should happen: (1) If a bitmap graphics is cut or copied, it should make a png copy to the X clipboard. (2) If it is a font, copy it to the X clipboard as a TTf. This is in accordance with the Free Desktop standard. Please refer to this page: http://www.freedesktop.org/wiki/Standards_2fClipboardsWiki Other data types should follow as the standard evolves. I think users expect this. For example I can copy a bitmap graphics in Konqueror and paste it into OpenOffice. Yes, here there is still much to do. The strange thing is that up to now people only rarely asked for this. Strange indeed since DND is used widely in NeXTStep. :-) It almost is there :-) Really, but we need to improve on it. I think the problem stems from incorrect Postscript font names in the generated ps data. I have tested it out and Ghostscript chokes on them. Regards, Charles -- panic(aha1740.c); /* Goodbye */ linux-2.2.16/drivers/scsi/aha1740.c RFC3156 defines security multipart formats for MIME with OpenPGP. pgp6VQewOJuRl.pgp Description: PGP signature ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: [Fwd: Re: really attracting developers]
IMHO there are two spaces which can be explored if you want to advance the GNUstep community: a) reasonably easy and convenient Windows porting for Cocoa developers b) server stuff I totally agree upon a.) This would definitely attract some Cocoa developers. Probably they would be mostly professionals. Which is not bad at all. However there's a point c). c) Users Users who really like the environment and use it on a daily basis tend to become developers. And to speak from my own experience: Users tend to like beautiful, customizable environments. They just need the possibility to kill some time doing useless things like playing with themes etc. Ciao Marc ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On Fri, 25 Aug 2006 12:05:48 +0200, Rogelio Serrano [EMAIL PROTECTED] wrote: Whats keeping other developers from gnustep? One thing might be that GNUStep-related stuff is somewhat disorganized. Documentation, for example, is scattered across the normal homepage, the wiki, http://gnustep.made-it.com/ and maybe others, and is sometimes even incomplete. The wiki feels kind of empty, underattended and stump-y. The application page references the GWorkspace guide but not the one for GNUMail (http://gnustep.made-it.com/Guides/GNUmail.html). There's no quick link for SVN access in the developers section, something that stumps me every time. Bug reports are stored at savannah.gnu.org, including the old but still intact CVS repo, while the new SVN repo is at gna.org. And so on. Some of these might be minor annoyances, but they add up, at least for me. Why not move everything, except the SVN repo and bug reports, to the wiki and make that the actual homepage? Additions, updates to, and discussions of guides and other information would be greatly eased, since people other than the maintainers can contribute quickly. A very quick look at the build guide suggests that it could be split up nicely to fit in. Snapshots of the documentation could then be included in releases and with some scripting work, you could possibly synchronize the API documentation in the wiki with the source and maybe vice-versa. Plus, a wiki is searchable without Google. Oh, while I'm day-dreaming about documentation, Yavor Doganov brought up Project Mallard (http://www.gnome.org/~shaunm/quack/mallard.xml) on the GNUMail ML. This would get along nicely with wikis me thinks. ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
Hi, On 25.08.2006, at 12:05, Rogelio Serrano wrote: Whats keeping other developers from gnustep? incomplete ide? incomplete nextstep based system? incomplete libraries? i don't buy the general applications unavailability argument. we are talking about people who want to create apps under gnustep. i don't buy the appearance argument either. It's funny (and sad...) that this discussion pops up every one and then since years ... IMO it is because some GNUstep developers don't want to see GNUstep being a complete desktop solution (for X11 based systems), instead they favour the cross-platform API path, whereas others don't want to write major applications for GNUstep because it is not a complete desktop and thus has no target audience - sadly GNUstep apps just don't integrate with other desktops. It is kind of a chicken and egg situation (yes I am aware of Etoile, but IMO I doubt this project will have a big impact on GNUstep in terms of attracting new users or developers - unless it gets integrated into GNUstep itself). This issue has never been resolved and as a consequence the project does not really make progress in terms of attracting new users or developers. Worse, while Java, .NET(Mono) or even gtk and Qt made progress in terms of installed base, functionality and maturity, GNUstep simply did not. Even more worse, you cannot use GNUstep to deploy OS X code on other OSes since the beasts are too different, e.g. there are many more APIs on OSX which are being used other than Cocoa when developing OS X apps. Not even the Objective-C runtime is 100% compatible (which will become a real issue with Leopard, I am afraid). So tell me one reason why I should write code for/with GNUstep unless I am an Objective-C aficionado .. (which I am, actually :). IMO GNUstep needs to change its focus otherwise it will always remain one of those many nice toolkits which will - unfortunately - never make it to the real surface... just my $0.02... and now hit me ;-) -Phil -- Philippe C.D. Robert http://www.nice.ch/~phip ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 2006-08-26 21:47:02 +0200 Philippe C.D. Robert [EMAIL PROTECTED] wrote: [... some very good points ...] just my $0.02... and now hit me ;-) Seconded -- Chris ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
Re: really attracting developers
On 2006-08-26 15:47:02 -0400 Philippe C.D. Robert [EMAIL PROTECTED] wrote: IMO it is because some GNUstep developers don't want to see GNUstep being a complete desktop solution (for X11 based systems), instead they favour the cross-platform API path, whereas others don't want to write major applications for GNUstep because it is not a complete desktop and thus has no target audience - sadly GNUstep apps just don't integrate with other desktops. I agree with this. The show stoppers that I found with GNUstep are: (1) A paste board that play nicely with the native clipboard system. (2) Drap and drop that works with the native system. (3) A print system that works. Charles -- Nature abhors a Vacuum -- Brian Behlendorf on OSS (Open Sources, 1999 O'Reilly and Associates) RFC3156 defines security multipart formats for MIME with OpenPGP. pgpENznBX9iGv.pgp Description: PGP signature ___ Discuss-gnustep mailing list Discuss-gnustep@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnustep
