Re: [Freedos-devel] Hello!
I think the original goal of FreeDOS has been met, based upon what I remember from back in 2000 or so when I came across the project. I have seen many posts back and forth about adding support for this and that because Windows sucks and so forth and so on. With that being said, here's what I see. Now I know opinions are like ... and everyone has one... The core FreeDOS aims to maintain compatibility with MS/PC DOS as it pertains to in-memory data structures, like SysVars, Job File Tables, memory control blocks, etc, documented (and undocumented API) based on DOS 6.22 such as the DOS routines (INT 21h), the multiplexer (INT 2FH), Ctrl-Break (INT 23H), Critical Error handler (INT 24H), and although superceded, absolute disk read and write (INT 25H, INT26H respectively). DOS was designed in an extensible manner, so I believe rather than changing the core of the OS, we need to EXTEND the OS. Obviously features such as IPv4/IPv6, GPT, are not relevant to the class of users that are wishing to still relive the glory days of the early x86 processors, i.e., 8086, 8086, 80186, 80286, etc. There is a different class of user, mostly those commenting on this thread, which would like to use DOS with newer machines. One alternative would be to design and build a protected mode kernel that would map all of the real mode calls and virtualize all hardware access. That would put us right where OSes like Linux and Windows are. The other alternative is to develop a platform that sits on top of DOS that runs and switches to protected mode and virtualize all the hardware. That puts us where Windows 95/98 was. To bring DOS into "the future" requires some parting with older technologies, which isn't particularly a goal of this project. In comparison, the issues that Microsoft has (and had) with Windows was in part due to their attempts to bring along the past into the future. This is why Microsoft let go of DOS in the Windows 9x code base and shifted to the Windows NT code base. Even then, with 32-bit code, it was still possible to run DOS applications (to an extent) but they were isolated to their own VM. Basically, short of forking this project, I don't see a way to incorporate the advanced features and still remain compatible with an OS that is over 20 years old. On Sun, May 17, 2015 at 6:07 PM JK Benedict wrote: > Excellent. > > "Daily" builds was just an example and I agree, daily builds would be > overkill. > > The angle I was coming from is when "core changes" start to be made. How > will this affect the 100 packages when core, resource, and drivers are > re-tooled? DOS is heavily classic, solid... but some of the changes will > affect the core. I should have been more specific in unit testing and > planning as various drivers, kernel, and kernel-deps change. > > Ah, Zip files -- the beauty of DOS. I love Linux, but sometimes I just > don't feel like writing code or compiling things :) > > Thanks for the reply and will research/download the wifi drivers as soon as > possible. > > --jesse/jkbs > > -Original Message- > From: Eric Auer [mailto:e.a...@jpberlin.de] > Sent: Sunday, May 17, 2015 7:17 AM > To: freedos-devel@lists.sourceforge.net > Subject: Re: [Freedos-devel] Hello! > > > Hi Jesse, > > Centralized documentation makes sense, but why would you put 100 packages > in > a centralized source code repository if 95 of them have not a single source > code change in a whole year? > > And why do nightly builds of all 100 then? DOS heavily relies on classic > software that simply is okay as it is and that no longer changes :-) > > As mentioned in the thread, there already is a considerable number of text > and graphical web browsers. It probably is better to improve one of those > instead of writing yet another browser. > > I agree that it is good to have a wishlist for shareware software that we > would like to become free open source. Maybe the list could be done in wiki > style? > > In general, if the hardware common for virtual machines is among the > hardware for which there are drivers, there is no need to have separate > development for virtualization and installation. > > We do already have a few VM-specific tools which are available :-) And > there > could be a download of a pre-installed VM, in case installation from ISO > takes too much effort ;-) > > IPv6 is widely available already but is rarely required so I agree that DOS > is not in a hurry. > > Regarding GPT, that is something that only needs some reasonably small > amount of kernel code to support in passive scenarios. Having FDISK with > GPT > would be way more code, I guess. Most other tools never look at a partition > table, so for them, this is not relevant. > > FileMaven basically does the LapLink thing, but it is closed source. It > would be nice to have something open. On computers with network (LAN), it > is > better to use existing FTP, SCP, SMB or HTTP tools to copy files around. > And > there is a tool to copy files between VM
Re: [Freedos-devel] Hello!
Excellent. "Daily" builds was just an example and I agree, daily builds would be overkill. The angle I was coming from is when "core changes" start to be made. How will this affect the 100 packages when core, resource, and drivers are re-tooled? DOS is heavily classic, solid... but some of the changes will affect the core. I should have been more specific in unit testing and planning as various drivers, kernel, and kernel-deps change. Ah, Zip files -- the beauty of DOS. I love Linux, but sometimes I just don't feel like writing code or compiling things :) Thanks for the reply and will research/download the wifi drivers as soon as possible. --jesse/jkbs -Original Message- From: Eric Auer [mailto:e.a...@jpberlin.de] Sent: Sunday, May 17, 2015 7:17 AM To: freedos-devel@lists.sourceforge.net Subject: Re: [Freedos-devel] Hello! Hi Jesse, Centralized documentation makes sense, but why would you put 100 packages in a centralized source code repository if 95 of them have not a single source code change in a whole year? And why do nightly builds of all 100 then? DOS heavily relies on classic software that simply is okay as it is and that no longer changes :-) As mentioned in the thread, there already is a considerable number of text and graphical web browsers. It probably is better to improve one of those instead of writing yet another browser. I agree that it is good to have a wishlist for shareware software that we would like to become free open source. Maybe the list could be done in wiki style? In general, if the hardware common for virtual machines is among the hardware for which there are drivers, there is no need to have separate development for virtualization and installation. We do already have a few VM-specific tools which are available :-) And there could be a download of a pre-installed VM, in case installation from ISO takes too much effort ;-) IPv6 is widely available already but is rarely required so I agree that DOS is not in a hurry. Regarding GPT, that is something that only needs some reasonably small amount of kernel code to support in passive scenarios. Having FDISK with GPT would be way more code, I guess. Most other tools never look at a partition table, so for them, this is not relevant. FileMaven basically does the LapLink thing, but it is closed source. It would be nice to have something open. On computers with network (LAN), it is better to use existing FTP, SCP, SMB or HTTP tools to copy files around. And there is a tool to copy files between VM and hypervisor. As mentioned elsewhere in this thread, there already are quite a few network drivers for DOS, but almost none for wireless network. Note that even if you do support the card, security protocols would still need a (often very complex) driver as well. Actually I agree with Mateusz: Better use a cheap portable and versatile access point with LAN between AP and DOS, so all the wireless complexity can be done by a small AP. There already is a FreeDOS repository of pre- packaged pre-compiled software that can be installed, both from file and over the network. Mateusz would be happy if you can help him to update and extend the contents. That repository also contains pre-packaged ZIPs with package sources. Remember that 95 out of 100 DOS tools do NOT get updated, so the sources are static and it works just fine to offer a ZIP with them for download. Cheers, Eric -- One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y ___ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel -- One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y ___ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
Re: [Freedos-devel] Hello!
Completely agree as the "web" today reveals two things: - How heavy it is - How bloated it is --jkbs -Original Message- From: Louis Santillan [mailto:lpsan...@gmail.com] Sent: Sunday, May 17, 2015 1:52 AM To: Technical discussion and questions for FreeDOS developers. Subject: Re: [Freedos-devel] Hello! On Sat, May 16, 2015 at 7:44 PM, JK Benedict wrote: [SNIP] > - Base resources, such as file system options/changes > - Connectivity tools > * Modernized web browser (I am working on one now - a text > based > prototype) There's Georg Potthast's Dillo port (in XFDOS [0]), and Arachne (besides links or lynx), both leave a lot to be desired in a JS heavy web. I personally think it could more productive to go after app specific tools for Google [1], Facebook [2], Twitter [3], LinkedIn [4], etc. For example, Facebook is entirely accessible over HTTP [5]. > - LapLink is still purchase-ware > * Has a specific protocol via LPT ports, etc > * PPP/PPPD in linux is open source... possible use of that? > * If not... let's write fdlink :) David Dunfield has LAPTALK, DDLINK & RDC [6]. If LAPTALK (source in MC323EXA) isn't sufficient, maybe he'd be willing to open source DDLINK and/or RDC? > * e1000 drivers (Gigabit)? 10/100? Tom Ehlert [7] & Georg Potthast [8] have gigabit NIC drivers. Dave Dunfield [9] has a nice collections as well. Georg & Dave also have nice NIC detection utils. [0] https://code.google.com/p/nanox-microwindows-nxlib-fltk-for-dos/wiki/XFDOS [1] https://youtu.be/yu1w1pPXz3Q [2] https://youtu.be/xrYRH3PYYT0 [3] https://youtu.be/FCvidD5JKBg [4] https://youtu.be/csuHU8B6Gg0 [5] https://developers.facebook.com/docs/graph-api/overview [6] http://www.classiccmp.org/dunfield/dos/index.htm [7] http://www.drivesnapshot.de/en/imakebootdisk.htm [8] http://www.georgpotthast.de/sioux/packet.htm [9] http://www.classiccmp.org/dunfield/dos/pktdrv.zip -- One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y ___ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel -- One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y ___ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
Re: [Freedos-devel] Hello!
Hi Jesse, Centralized documentation makes sense, but why would you put 100 packages in a centralized source code repository if 95 of them have not a single source code change in a whole year? And why do nightly builds of all 100 then? DOS heavily relies on classic software that simply is okay as it is and that no longer changes :-) As mentioned in the thread, there already is a considerable number of text and graphical web browsers. It probably is better to improve one of those instead of writing yet another browser. I agree that it is good to have a wishlist for shareware software that we would like to become free open source. Maybe the list could be done in wiki style? In general, if the hardware common for virtual machines is among the hardware for which there are drivers, there is no need to have separate development for virtualization and installation. We do already have a few VM-specific tools which are available :-) And there could be a download of a pre-installed VM, in case installation from ISO takes too much effort ;-) IPv6 is widely available already but is rarely required so I agree that DOS is not in a hurry. Regarding GPT, that is something that only needs some reasonably small amount of kernel code to support in passive scenarios. Having FDISK with GPT would be way more code, I guess. Most other tools never look at a partition table, so for them, this is not relevant. FileMaven basically does the LapLink thing, but it is closed source. It would be nice to have something open. On computers with network (LAN), it is better to use existing FTP, SCP, SMB or HTTP tools to copy files around. And there is a tool to copy files between VM and hypervisor. As mentioned elsewhere in this thread, there already are quite a few network drivers for DOS, but almost none for wireless network. Note that even if you do support the card, security protocols would still need a (often very complex) driver as well. Actually I agree with Mateusz: Better use a cheap portable and versatile access point with LAN between AP and DOS, so all the wireless complexity can be done by a small AP. There already is a FreeDOS repository of pre- packaged pre-compiled software that can be installed, both from file and over the network. Mateusz would be happy if you can help him to update and extend the contents. That repository also contains pre-packaged ZIPs with package sources. Remember that 95 out of 100 DOS tools do NOT get updated, so the sources are static and it works just fine to offer a ZIP with them for download. Cheers, Eric -- One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y ___ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
Re: [Freedos-devel] Hello!
There is WIFI drivers available,but they are often made for a specific wireless card.Your best bet is to make a driver that can interpret ethernet signals.Since information is coming from the internet to the ethernet cables,you can get information/data from the internet using a parallel port driver.This may help. -Jayden On Sun, May 17, 2015 at 1:03 AM, Steve Nickolas wrote: > On Sat, 16 May 2015, Louis Santillan wrote: > > > On Sat, May 16, 2015 at 7:44 PM, JK Benedict > wrote: > > [SNIP] > >> - Base resources, such as file system options/changes > >> - Connectivity tools > >> * Modernized web browser (I am working on one now - a text based > >> prototype) > > > > There's Georg Potthast's Dillo port (in XFDOS [0]), and Arachne > > (besides links or lynx), both leave a lot to be desired in a JS heavy > > web. I personally think it could more productive to go after app > > specific tools for Google [1], Facebook [2], Twitter [3], LinkedIn > > [4], etc. For example, Facebook is entirely accessible over HTTP [5]. > > I agree. An all-purpose browser needs code to handle a lot of corner case > stuff, but a one-purpose browser only needs to be able to handle the > specific case it was designed for. (But it may need to be able to update > itself in case an update to the site breaks it.) > > I have considered writing such tools, but my head exploded at the idea of > having to do so much string parsing in C, a language that really can't > handle it well. To date the only apps I managed to complete were rdate > and airc (both using wattcp), and the source code to the former is now > lost. > > (Keep in mind my target system was a VERY low resource environment, such > as the 5160 I still have in my closet.) > > >> - LapLink is still purchase-ware > >> * Has a specific protocol via LPT ports, etc > >> * PPP/PPPD in linux is open source... possible use of that? > >> * If not... let's write fdlink :) > > > > David Dunfield has LAPTALK, DDLINK & RDC [6]. If LAPTALK (source in > > MC323EXA) isn't sufficient, maybe he'd be willing to open source > > DDLINK and/or RDC? > > I almost would want to suggest that a FreeDOS app for this purpose mimic > INTERLNK/INTERSVR, since that was such a tool actually included in PC DOS > 5.02 and later, and MS-DOS 6.0 and later. (Some old-timers here might > know I'm a little too familiar with MS and PC DOS and what each version > comes with. :P That's actually why I no longer actively contribute.) > > As for PPP, you already have LSPPP as a GPL'd option. If it works, and I > believe it does implement the Crynwr packet driver interface most TCP apps > for DOS require, there's no need to reinvent the wheel. > > http://ladsoft.tripod.com/lsppp.html > > -uso. > > > -- > One dashboard for servers and applications across Physical-Virtual-Cloud > Widest out-of-the-box monitoring support with 50+ applications > Performance metrics, stats and reports that give you Actionable Insights > Deep dive visibility with transaction tracing using APM Insight. > http://ad.doubleclick.net/ddm/clk/290420510;117567292;y > ___ > Freedos-devel mailing list > Freedos-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/freedos-devel > -- One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y___ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
Re: [Freedos-devel] Hello!
On Sat, 16 May 2015, Louis Santillan wrote: > On Sat, May 16, 2015 at 7:44 PM, JK Benedict wrote: > [SNIP] >> - Base resources, such as file system options/changes >> - Connectivity tools >> * Modernized web browser (I am working on one now - a text based >> prototype) > > There's Georg Potthast's Dillo port (in XFDOS [0]), and Arachne > (besides links or lynx), both leave a lot to be desired in a JS heavy > web. I personally think it could more productive to go after app > specific tools for Google [1], Facebook [2], Twitter [3], LinkedIn > [4], etc. For example, Facebook is entirely accessible over HTTP [5]. I agree. An all-purpose browser needs code to handle a lot of corner case stuff, but a one-purpose browser only needs to be able to handle the specific case it was designed for. (But it may need to be able to update itself in case an update to the site breaks it.) I have considered writing such tools, but my head exploded at the idea of having to do so much string parsing in C, a language that really can't handle it well. To date the only apps I managed to complete were rdate and airc (both using wattcp), and the source code to the former is now lost. (Keep in mind my target system was a VERY low resource environment, such as the 5160 I still have in my closet.) >> - LapLink is still purchase-ware >> * Has a specific protocol via LPT ports, etc >> * PPP/PPPD in linux is open source... possible use of that? >> * If not... let's write fdlink :) > > David Dunfield has LAPTALK, DDLINK & RDC [6]. If LAPTALK (source in > MC323EXA) isn't sufficient, maybe he'd be willing to open source > DDLINK and/or RDC? I almost would want to suggest that a FreeDOS app for this purpose mimic INTERLNK/INTERSVR, since that was such a tool actually included in PC DOS 5.02 and later, and MS-DOS 6.0 and later. (Some old-timers here might know I'm a little too familiar with MS and PC DOS and what each version comes with. :P That's actually why I no longer actively contribute.) As for PPP, you already have LSPPP as a GPL'd option. If it works, and I believe it does implement the Crynwr packet driver interface most TCP apps for DOS require, there's no need to reinvent the wheel. http://ladsoft.tripod.com/lsppp.html -uso. -- One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y ___ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
Re: [Freedos-devel] Hello!
Hello JK, That's an extensive list :) On 17/05/2015 04:44, JK Benedict wrote: > Priority 1: Centralized Documentation > ~~ > - Single-source for links to mailing list, project goals, components, > completion, owners, priority, etc This is the purpose of www.freedos.org I believe. > - Current packages Here I think FDNPKG should be used as the reference, as it is supposed to contain the latest versions of everything: http://www.ibiblio.org/pub/micro/pc-stuff/freedos/files/distributions/1.1/repos/listing.txt > Priority 3: Modernization > > - Network and Modems > * So, who has spare hardware? > * I can write modem utilities, but I need the modems... trashed all > mine 2-3 years ago and only have a Roboticks 14.4 I don't really think writing drivers for modems would be very usefull today, I am pretty sure almost nobody use a modem to access the internet nowadays. Most people have a SOHO router that they connect to their machines either via an ethernet cable or wifi (and here FreeDOS performs very poorly, since we have no useable wifi drivers). > Any thoughts to a FreeDOS "repo"? One that can be downloaded entirely OR > via the FreeDOS command-line? Would be kinda cool! I know wget is already > there, but having a util to pull from source w/ manifesto would help speed > up and track/keep a database of things to be installed, tested, uninstalled, > etc. That's what FDNPKG is about. But maybe you are referring to some other kind of repo, not a repository with installable packages? Mateusz -- One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y ___ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel