Apologies. I resend this hoping it will be more readable this time. 
Unfortunately this looks good on my side when I send it. So I do initially not 
know what goes wrong.

On 27/06/2013, at 17:28 , Christiaan Hofman wrote:


On Jun 27, 2013, at 16:33, Fischlin Andreas wrote:



On 27/06/2013, at 13:08 , Christiaan Hofman wrote:


On Jun 27, 2013, at 12:50, Fischlin Andreas wrote:

Dear Christiaan,

Ok. I understand now finally, sic. I was able to create from existing bookmarks 
the needed new bookmarks, which indeed do now the job.

I guess then the problem is (once more) with the lack of an editor for the 
server settings or perhaps all settings of an existing search bookmark.

My conclusions:  I can do what I needed to do, but there remains an issue. I 
find this edit cycle again - I remember now, having stumbled previously over 
this BD design issue -  not particularly user friendly, at least idiosyncratic, 
since not logical. Having reflected all what you said, then I'd say it is a 
general issue of BD that bookmarks can only be edited by using an existing 
bookmark (via choosing its menu), then editing that external group, then saving 
the edited group as a bookmark, and then discarding the original bookmark and 
replacing it by the newly edited/saved one.

IMHO I find it hard to get ones head around this. If there would be a sheet to 
edit an existing search bookmark or perhaps any server settings, I guess I 
would never have had these difficulties to figure out how to actually 
accomplish a simple task such as setting preferences for a particular search 
group. Perhaps something you could consider. It may be very clear to you as a 
chief programmer of the GUI what steps need to be done to accomplish this task, 
but there is the danger with all this knowledge to forget about those like me, 
who do not share the same intimate knowledge with the internals of the software.

In all software design I prefer to distinguish clearly between the user and the 
implementer's model and strive always to keep the user model as simple as 
possible and as user-friendly as possible. As a user of BD I know that there 
are some objects like "Search groups", "Data base server settings", "Search 
criteria", and then "Search bookmarks" with some relationships among them. For 
instance the latter consisting of the other elements, some of which are 
mandatory ("Data base server settings"), some optional (search criteria).

So why not having another menu command to edit existing "Search bookmarks" 
which behaves similarly to the "New Search Group..." menu command, but which 
operates on an existing instantiation? Alternatively the menu command "Edit 
Search Bookmarks..." could of course also do that, if its dialog would be 
extended by being able to Ctrl-click a line and/or having an additional tool in 
the top toolbar that triggers the "Get Info for Group..." sheet. Then no new 
menu command would be needed.

If any of these features would exist in BD, then I guess, all our discussions 
would never have taken place.

Please consider. Otherwise thanks once more for all what you do.

Regards,
Andreas


The problem is that the model is complex. The server information is complex, 
and therefore the UI for it is complex. and then there are the various things 
that can be edited, that make it complex. Adding more interfaces does not 
necessarily make things simpler, it complicates things even further. Your first 
suggestion is certainly a no-op, it confuses more rather than less (you could 
then edit 3 things rather than 2 in the same place: a group instance, a default 
server, or a bookmark) I also doubt that the discussion would have been avoided 
by adding new UI in a different place. Why would it?

I seem to have a slightly different view, since I guess for the reasons I 
already gave. But I understand that this is complex and I am also for keeping 
it as simple as possible.


In general, a bookmark is simply a static template. Always. When you choose it, 
you create an instance. I know of no app where changing that instance would 
change the static template, that makes no sense. For instance, when you choose 
a bookmark in a browser, and then change the address for the page, why would 
you expect the bookmark to change?

Good example, yes I do exactly not expect it to change. However, let's not 
confound oranges with apples. I draw another conclusion from this. When I have 
an editor for bookmarks, I expect it to be able to edit the address of a 
bookmark when for instance the provider did change the web site. E.g. when I 
wish to keep a bookmark menu for BD's home page and the address of the latter 
changes, then I find it cumbersome having to first visit the new home page by 
choosing that bookmark menu (which does in this case no longer work) merely to 
preserve some of the previous settings of that bookmark, then having to save 
the new URL once successfully visited the new home page as a new bookmark, and 
then having to delete the old bookmark after having moved the new bookmark to 
the same command location where the old bookmark menu was. When I have an 
editor for bookmarks, as offered by all browsers AFAIK, then I expect to be 
able to edit the address of that bookmark directly, e.g. by adding credenti
 als (our case) or editing the URL string otherwise. However, with the current 
BD GUI this is not possible and that's what I believe is so difficult to 
understand from the user's point of view.


This is a request for a new feature. That is really a separate subject from 
what we're discussing, not directly related to the password/username but 
general editing of searches.

Agreed.


Moreover, I am not talking about more interfaces. I was thinking of using the 
same routine (edit sheet) for editing an existing instantiation as is used for 
editing a new instantiation. In my programmer's mind this simply means another 
argument for the same algorithm/routine/method, where properties are not empty 
(or predefined), but the last set ones for the instantiation being the argument.

You can edit an existing group. However, that starts with the group as it is. 
It makes no sense to start from whatever you did last. Or if with 
"instantiations" you mean bookmarks (it's not what I would think of), then it 
is in fact new UI, even if it reuses old ones. But it's a separate, and 
different, question than what we're really talking about.

yes, see above.



Then I believe you have misunderstood me. I do not expect a static template to 
be changeable, of course not. That's exactly one of the reasons why I find the 
current solution so confusing, with having to start from an instantiation when 
one wishes to edit some properties of an entire class. Not only a long detour, 
but also a confusing one.


The thing is that there really isn't something as a class. That's what you seem 
to misunderstand. There are just various presets. And you can edit or add new 
presets. This is because it's pretty versatile.

In a user model there are objects, almost always, regardless whether you are 
conscious of that fact or not. And I believe that's not what you are referring 
to. But let's stop this discussion, since it goes way beyond the issue where we 
started and I agree that your solution offers much more flexibility than the 
one with the hidden preferences and it seems to work well according to my first 
tests. It only is quite "hidden" and probably anyway quite out of reach for 
many users.

Therefore most of all, many thanks for your contribution (and to Colin) for 
making Web Of Science searching with BibDesk again possible. Thanks!!!!!!!  
:-)))))))


Again, I believe, only arguments that speak for aforementioned editing feature 
that can modify the properties of an existing bookmark.

But I am not hang up on this. I simply believe this needs improvement if 
possible, since current implementation is not sufficiently user model oriented.

Regards,
Andreas

Maybe. But asking for an extra bookmark edit feature is a separate issue.

Agreed

Cheers,
Andreas


Christiaan


Christiaan


On 27/06/2013, at 11:26 , Christiaan Hofman wrote:


On Jun 27, 2013, at 10:19, Fischlin Andreas wrote:

Dear Christiaan,

snip ...

I found the possibility to enter for External Group "Web of Science SCI" via 
Ctrl-Click "Get Info for Group..." a sheet where I could enter UserID and 
password. Is this what you mean?


Yes.

Ok. Good. I understand now the way how you have implemented that. This is of 
course also quite a good solution, if it only would work (see below).


However, this seems to be a most cumbersome method, since it seems I have to 
reenter all the time the UserID and the password upon each reuse of another ISI 
WOS Search (I have committed the changes using 'Set'), since the display of the 
two field's contents changes (I can't test this, unless I go home where I have 
no IP# authentication interference). Is this really what you intend to ask from 
users?


You should only have to set it once. You can set the changes for the default 
group type, or create a separate default group type with the password. It 
doesn't change when you've saved the changes, it certainly doesn't for me. So 
you must have done something wrong if it changes for you.


What could be the reason for the fact that it does not work? How can I set it 
globally and once and for all? Let me ask you once more whether what I do is 
really the technique intended. This is what I see as the only possibility to 
enter the credentials:

1) I choose BD menu command "Searches -> Web of Science SCI"
2) I Ctrl-Click on the resulting external group "Web of Science SCI" near the 
top left pane
3) I choose from pop-up menu "Get Info for Group..."
4) I click on arrow (top right corner) to cause the sheet for server settings 
to come down
5) I click onto button "Add"
6) I click onto button "Edit"
7) I dismiss the instructional dialog re using button "Set" by clicking button 
"OK"
8) I enter Username and Password
9) I commit the editing by clicking button "Set"
10) I dismiss the sheet by clicking button "OK"
11) I can finally enter a search criteria and cause a search to take place

Up to here all fine, although I wonder how editing of a specific group should 
affect all groups of that kind.

It doesn't. It changes the item in the default servers popup.


12) I repeat step 1)
13) I enter another search criteria

Search fails due to authentication failure and indeed,

14) repeating steps 2) and 3) shows that no UserID nor Password is set for that 
second external group "Web of Science SCI" empty, only the first group for 
which I have gone through all these steps 3) to 10) can I successfully 
authenticate.

You did not change the item in the search bookmarks menu. You should save that 
separately. The bookmarks are an already fully set up group. You could have 
also saved a setup group with password on the search bookmarks, rather than the 
default server popup. If you had done "New Search Group" and then chosen the 
item from the popup you would get the username and password.


This is all unacceptably cumbersome.

It's in principle a one-time setup. So I don't agree. You have to pass the 
information somehow.

So I ask again, is this really the intended behavior?

Yes.

If that should be the case then I opt clearly to go back to Colin's solution 
with hidden preferences.

Hidden prefs are bad. There's nothing in the app to tell the user it even 
exists.

Security is no issue here, since you also store the password not in the 
keychain.

If this is not the intended behavior as I hope is actually the case, then there 
is a bug in the implementation.

No, a misunderstanding on your side. You did not understand that the 3 buttons 
in the sheet refer to the default servers, i.e. the ones in the popup, as the 
label indicates. And bookmarks are specific instances of search groups.

I guess it has to do with the fact that the editing can only be done (at least 
AFAIK) for a specific instantiation of the external group "Web of Science SCI" 
instead of the class of these searches. Only then I presume all instantiations 
could inherit the needed credentials.

Please clarify.

Thanks for your help and cooperation.

Regards,
Andreas

It can be done for specific instants of search groups, *or* for default 
(preset) servers. But not for all instances of a generic type (there is no such 
information.)

Christiaan


------------------------------------------------------------------------------
This SF.net<http://SF.net/> email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Bibdesk-develop mailing list
[email protected]<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/bibdesk-develop


------------------------------------------------------------------------------
This SF.net<http://SF.net/> email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Bibdesk-develop mailing list
[email protected]<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/bibdesk-develop


------------------------------------------------------------------------------
This SF.net<http://SF.net/> email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Bibdesk-develop mailing list
[email protected]<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/bibdesk-develop


------------------------------------------------------------------------------
This SF.net<http://SF.net/> email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Bibdesk-develop mailing list
[email protected]<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/bibdesk-develop

------------------------------------------------------------------------------
This SF.net<http://SF.net/> email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev_______________________________________________
Bibdesk-develop mailing list
[email protected]<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/bibdesk-develop


ETH Zurich
Prof. Dr. Andreas Fischlin
Systems Ecology - Institute of Integrative Biology
CHN E 21.1
Universitaetstrasse 16
8092 Zurich
SWITZERLAND

[email protected]<mailto:[email protected]>
www.sysecol.ethz.ch<http://www.sysecol.ethz.ch>

+41 44 633-6090 phone
+41 44 633-1136 fax
+41 79 595-4050 mobile

             Make it as simple as possible, but distrust it!
________________________________________________________________________








------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Bibdesk-develop mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bibdesk-develop

Reply via email to