OS2 World.Com Forum

Subject  :  Voyager and taking over the world :-)
Author  :  cytan
Date  :  14 Apr, 2006 on 15:04
I just listened to the ogg and read the pdf presentations from the Developer workshop 2006 and found it absolutely interesting.

The news about migrating from the OS/2 kernal to a as yet undecided unix kernel is encouraging news. Keeping the WPS SOM model is clearly the way to go. I think the best part is that the new SOM will be done on OS/2 and the as yet to be chosen kernel giving us a migration path as well as encouraging other coders/developers to join in the effort. I think that we have a dearth of developers in the OS/2 community and for Voyager to actually deliver, more manpower is needed.

I would really like to encourage people here to download at least
Adrian's proposal and see how we can contribute to Voyager.

ftp://ftp.netlabs.org/pub/events/DWS2006/ogg/TheVoyagerProject-dws06.ogg

For OS/2 to survive, we need this project to work.

cytan


Subject  :  Re:Voyager and taking over the world :-)
Author  :  Sebadoh sys3175@optonline.net
Date  :  15 Apr, 2006 on 06:10

cytan (14 Apr, 2006 15:04):
I just listened to the ogg and read the pdf presentations from the Developer workshop 2006 and found it absolutely interesting.

The news about migrating from the OS/2 kernal to a as yet undecided unix kernel is encouraging news. Keeping the WPS SOM model is clearly the way to go. I think the best part is that the new SOM will be done on OS/2 and the as yet to be chosen kernel giving us a migration path as well as encouraging other coders/developers to join in the effort. I think that we have a dearth of developers in the OS/2 community and for Voyager to actually deliver, more manpower is needed.

I would really like to encourage people here to download at least
Adrian's proposal and see how we can contribute to Voyager.

ftp://ftp.netlabs.org/pub/events/DWS2006/ogg/TheVoyagerProject-dws06.ogg

For OS/2 to survive, we need this project to work.

cytan


It is a silly idea. OS/2 is over when 32bit processor (even in legacy) go away. The end. Stupid idea to devote the small amount of developers the OS/2 community has to a project which only strives to replace the only part of OS/2 which doesn't really need replacement at this point.


Subject  :  Re:Voyager and taking over the world :-)
Author  :  osw
Date  :  15 Apr, 2006 on 11:44

Sebadoh (15 Apr, 2006 06:10):

cytan (14 Apr, 2006 15:04):
I just listened to the ogg and read the pdf presentations from the Developer workshop 2006 and found it absolutely interesting.

The news about migrating from the OS/2 kernal to a as yet undecided unix kernel is encouraging news. Keeping the WPS SOM model is clearly the way to go. I think the best part is that the new SOM will be done on OS/2 and the as yet to be chosen kernel giving us a migration path as well as encouraging other coders/developers to join in the effort. I think that we have a dearth of developers in the OS/2 community and for Voyager to actually deliver, more manpower is needed.

I would really like to encourage people here to download at least
Adrian's proposal and see how we can contribute to Voyager.

ftp://ftp.netlabs.org/pub/events/DWS2006/ogg/TheVoyagerProject-dws06.ogg

For OS/2 to survive, we need this project to work.

cytan


It is a silly idea. OS/2 is over when 32bit processor (even in legacy) go away. The end. Stupid idea to devote the small amount of developers the OS/2 community has to a project which only strives to replace the only part of OS/2 which doesn't really need replacement at this point.



well, you comment made me worry sebadoh....
i'm visiting osnews.com reading a lot about some projects - not even half exciting as voyager - not even half modern and functionall as os/2 - but they do exist and draw attention... that's why I think your opinion is not fair - i'd love to use 64-bit os/2 (or os/3...) as well as wps with soft shadows and transparency, better fonts support and many things which won't happen on existing os/2 nor ecs just because code is closed by ibm
and not because it can not be rewriten.
If it can be rewriten it should be, and voyager is about to rewrite things (wps) again, to do something instead of begging ibm to free code... it's time to say "bye" to IBM and do something on our own.

But of course I can be wrong, and will change my mind if you (or someone else) propose and define (at least as good as voyager does) our needs, target and the way to replace entire system at once (i'd prefer that).

So to all talkers: let doers do their job.

bye/2


Subject  :  Re:Voyager and taking over the world :-)
Author  :  Shai none-brain@web.de
Date  :  15 Apr, 2006 on 15:19
If I understood it the right way, Voyager isn't going to be a kernel replacement based on Linux/Unix technology but a kind of OS/2-like Window manager and WPS based on Linux/Unix. The auther compares it to Mac OS X - or did they change their aims recently?
I think this maybe possible but I don't think it's good. I don't use OS/2 to switch to Linux someday. I don't like the basics of Linux at all. It's just like one big patchwork and it's a miracle it still works. Even and especially an open source product should be as well planned and as professional as any closed source program.
That's why I think we shouldn't build a new OS/2 upon Posix technology.
Why not write an OS/2 subsystem for ReactOS???
It would be much easier, we would have an up-to-date kernel and system and a clean windowing system - not based on Linux at all. And - even more importantly - we would have Windows 3.11 & 9x support from the first day. If we have a working OS/2 Api implementation, we could replace the Windows-like ReactOS explorer with a WPS rewrite (but of course it would work without, too).
I would even consider a more exotic possibility. Do you know the HXRT project? It's a 32-bit DOS extender that can load PE executables and provides a Win32 API for PLAIN DOS. You can run Win9x console applications, simple GPI programs, SDL and even OpenGL programs with it - Odin for DOS (Yes they ran Quake II under plain DOS with this little gadget.). Using the same technology (by maybe just extending HXRT), I bet we can run most OS/2 VIO programs in just a few weeks of work. Sounds weird, but DOS (particular FreeDOS 32-bit) IS an up-to-date operating system (32-bit kernel, up-to-date drivers, FAT-32 support...) and using this technology as a base would be a great easement. More importantly, it's a realistic possibility although I know you won't like it.
What do you think?

Robin


Subject  :  Re:Voyager and taking over the world :-)
Author  :  tex
Date  :  15 Apr, 2006 on 15:39

Shai (15 Apr, 2006 15:19):
If I understood it the right way, Voyager isn't going to be a kernel replacement based on Linux/Unix technology but a kind of OS/2-like Window manager and WPS based on Linux/Unix. The auther compares it to Mac OS X - or did they change their aims recently?
I think this maybe possible but I don't think it's good. I don't use OS/2 to switch to Linux someday. I don't like the basics of Linux at all. It's just like one big patchwork and it's a miracle it still works. Even and especially an open source product should be as well planned and as professional as any closed source program.
That's why I think we shouldn't build a new OS/2 upon Posix technology.
Why not write an OS/2 subsystem for ReactOS???
It would be much easier, we would have an up-to-date kernel and system and a clean windowing system - not based on Linux at all. And - even more importantly - we would have Windows 3.11 & 9x support from the first day. If we have a working OS/2 Api implementation, we could replace the Windows-like ReactOS explorer with a WPS rewrite (but of course it would work without, too).
I would even consider a more exotic possibility. Do you know the HXRT project? It's a 32-bit DOS extender that can load PE executables and provides a Win32 API for PLAIN DOS. You can run Win9x console applications, simple GPI programs, SDL and even OpenGL programs with it - Odin for DOS (Yes they ran Quake II under plain DOS with this little gadget.). Using the same technology (by maybe just extending HXRT), I bet we can run most OS/2 VIO programs in just a few weeks of work. Sounds weird, but DOS (particular FreeDOS 32-bit) IS an up-to-date operating system (32-bit kernel, up-to-date drivers, FAT-32 support...) and using this technology as a base would be a great easement. More importantly, it's a realistic possibility although I know you won't like it.
What do you think?

Robin


The kernel is eventually going to have to be replaced since Serenity does not have the source code for it. So what if the WPS sits on top of a different kernel so long as WPS does the same thing and does it better. In my opinion, what they have planned is good. It is not the kernel that sold me on OS/2 years ago, but the WPS. Even today, some of its features have not been duplicated.

David


Subject  :  Re:Voyager and taking over the world :-)
Author  :  RobertM
Date  :  15 Apr, 2006 on 16:16
The WPS is also one of my favorite parts of OS/2, but the kernel is pretty high up there as well. OS/2 handles and schedules threads, and handles & utilizes multiple CPUs/cores better than any Intel based OS out there. Duplicating that capability is going to be a challenge, though I think that task in good hands.

Also, HPFS/HPFS386 support would be a must for me. JFS is nice, but HPFS386 I've found to be better in numerous ways (with few exceptions, like maximum partition size and partition resizing).

Just my one cent worth,
-Robert


Subject  :  Re:Voyager and taking over the world :-)
Author  :  lpino
Date  :  15 Apr, 2006 on 18:30
The idea of replacing OS/2 with an open source project it's a very old one. At least from the year 2000 there has been people talking and working on and off on this. The greatest difficulty has been the agreement on what to use to replace the kernel. As someone said already the OS/2 kernel it's a piece of art when it comes to multitasking (under Intel). To me only BeOS was superior (real time was better)
WPS is one of the main selling points of OS/2. By far the most advance ideas on the desktop and replacing with an open source projects sounds great!!. They should start at the same level Warp 4 was based on SOM 2.1 and then move onto SOM 3. They should develop tools to create new code (right now only VAC 3.08 can create code for WPS) under GCC or Watcom.
I hope some day the project osFree (this is a project to replace the kernel with a microkernel based on L4, there is a Linux based on L4 and now Hurd is being develop using L4) can joint Netlabs and more developers can help, it could be taken as a student project or something too.
Another piece I like about OS/2 is the multimedia subsystem. I know is buggy as hell but the ideas behind it are very good ones. The IOPROC system is very flexible and helps with the whole WPS idea.
Please, people be positive about the work being done. There is no way everyone is going to be happy about the decisions taken but at least something is being done, so lets help with good spirit and if we don't like it and we are not doing anything to help OS/2 survive, we should be quiet and go out and put money into the bountys we do believe.

Leonardo Pino


Subject  :  Re:Voyager and taking over the world :-)
Author  :  osw
Date  :  15 Apr, 2006 on 19:13

Shai (15 Apr, 2006 15:19):
If I understood it the right way, Voyager isn't going to be a kernel replacement based on Linux/Unix technology but a kind of OS/2-like Window manager and WPS based on Linux/Unix. The auther compares it to Mac OS X - or did they change their aims recently?
I think this maybe possible but I don't think it's good. I don't use OS/2 to switch to Linux someday. I don't like the basics of Linux at all. It's just like one big patchwork and it's a miracle it still works. Even and especially an open source product should be as well planned and as professional as any closed source program.
That's why I think we shouldn't build a new OS/2 upon Posix technology.
Why not write an OS/2 subsystem for ReactOS???
It would be much easier, we would have an up-to-date kernel and system and a clean windowing system - not based on Linux at all. And - even more importantly - we would have Windows 3.11 & 9x support from the first day. If we have a working OS/2 Api implementation, we could replace the Windows-like ReactOS explorer with a WPS rewrite (but of course it would work without, too).
I would even consider a more exotic possibility. Do you know the HXRT project? It's a 32-bit DOS extender that can load PE executables and provides a Win32 API for PLAIN DOS. You can run Win9x console applications, simple GPI programs, SDL and even OpenGL programs with it - Odin for DOS (Yes they ran Quake II under plain DOS with this little gadget.). Using the same technology (by maybe just extending HXRT), I bet we can run most OS/2 VIO programs in just a few weeks of work. Sounds weird, but DOS (particular FreeDOS 32-bit) IS an up-to-date operating system (32-bit kernel, up-to-date drivers, FAT-32 support...) and using this technology as a base would be a great easement. More importantly, it's a realistic possibility although I know you won't like it.
What do you think?

Robin



wow! "Weird Ideas Contest 2006" is open! Hurrah!

32-bit systems with fat32 support were up to date about 10 years ago... we have 64-bits now as well as bunch of better filesystems there, even good old warp 3 was able to use fat32 not to mention 32-bits here. So what's the point to stick to dos? All things have their own right place in universe.... for dos (any kind of it) that place is called Museum.

And what about reactOS? Why I'm not happy? Wouldn't it be wonderfull to have some support for 16-bit text os/2 apps in box under windose NT clone (with questionable copyrights to part of it's code...) - just like 12 years ago in winnt 3.51? Was there ever plans to have support for wps apps under reactOS?
btw. will ms let to decrease it's incomes by free clone of its product? let's be serious...
the day when reactOS becomes usable anyway (if everl) will be also the day when ms lawyers start to work hard for their salaries

Instead of inventing round rectangles let's do free os/2 clone with free bugless wps and 64-bits under hood. Voyager is about to make better, nicer, free and open WPS. Osfree "was" to provide base for it (already choosen kernel - L4 - modern mach version) , lots of rewritten (opened) text apps - just ready to recompile on 64-bits... and of course no interest from os/2 community... Now I know why. Most of warp users would like to switch to dos or winNT at least.


Subject  :  Re:Voyager and taking over the world :-)
Author  :  Shai none-brain@web.de
Date  :  15 Apr, 2006 on 21:15

osw (15 Apr, 2006 19:13):
wow! "Weird Ideas Contest 2006" is open! Hurrah!


Thanks for your pessimism.

32-bit systems with fat32 support were up to date about 10 years ago... we have 64-bits now as well as bunch of better filesystems there, even good old warp 3 was able to use fat32 not to mention 32-bits here. So what's the point to stick to dos? All things have their own right place in universe.... for dos (any kind of it) that place is called Museum.

It's simply one possibility if you consider 32-bit operating systems. I admit it's really not the best one but it would be relatively easy to have usable results on open source basis quickly.
By the way: If every 32-bit based operating system has its place in a museum then OS/2 already belongs in a museum, too? DOS has its place outside a museum because of its simplicity and transparency.

And what about reactOS? Why I'm not happy? Wouldn't it be wonderfull to have some support for 16-bit text os/2 apps in box under windose NT clone (with questionable copyrights to part of it's code...) - just like 12 years ago in winnt 3.51? Was there ever plans to have support for wps apps under reactOS?
btw. will ms let to decrease it's incomes by free clone of its product? let's be serious...

That's really not the same. At first ReactOS isn't a simple "NT clone" but a Win9x compatible operating system. OS/2 support could work through a subsystem, more or less equal to the Win9x part. Why should it be impossible to support PM and 32-bit OS/2 applications (which of course should be the main focus)?? WPS is just an user interface - it can be added to every (gui) OS. (There may be even some comercial WPS implementations for Windows that already work with ReactOS. According to the ReactOS homepage there are already some people who considered writing a subsystem for OS/2. Does anyone know more about such efforts?
ReactOS probaply isn't based on MS code at all. They were just blamed for it. Currently they completely check their code to eliminate this rumor. This process is completed 70% as of now.

the day when reactOS becomes usable anyway (if everl) will be also the day when ms lawyers start to work hard for their salaries

It's already usable - not for the everyday work and not for the public but they got quite far yet. Just look at the screenshots section of their homepage (http://www.reactos.org/xhtml/en/screenshots.html)
I have the recent version running in VirtualPC but transferring files to the system is difficult (no virtual network drive support; I use the LiveCD because a installed ReactOS system works not so good with VirtualPC), so I just tried some programs. FORTRAN/TK programs run (although they are sloooow) *g*.

Instead of inventing round rectangles let's do free os/2 clone with free bugless wps and 64-bits under hood. Voyager is about to make better, nicer, free and open WPS. Osfree "was" to provide base for it (already choosen kernel - L4 - modern mach version) , lots of rewritten (opened) text apps - just ready to recompile on 64-bits... and of course no interest from os/2 community...

You know that Voyager won't be just an OS/2 compatible WPS replacement but - as far as I know - a windowing system and WPS-clone for Linux/Unix. It probaply won't be OS/2 (backwards) compatible at all. It would be similar to GNOME as it was once planned but without X11.
OS/2 compatibility is in my opinion a precondition for an "OS/2 clone". That's why I'm really not against osFree - it's a clean attempt. But building an operating system nearly from scratch is quite difficult and lengthy - maybe too difficult for our weak community. I just wanted to show the possibility to base an OS/2 clone on an existing operating system (and a non-posix one unlike Voyager).
If so many people just use OS/2 because of the WPS, we could write a WPS for no-matter-what operating system. If it has to be open source, why not just another X window manager?

Now I know why. Most of warp users would like to switch to dos or winNT at least.

This was my personal opinion - and they were just examples/possibilities.

Robin


Subject  :  Re:Voyager and taking over the world :-)
Author  :  cytan
Date  :  16 Apr, 2006 on 04:23
From what I understand the kernel has not been chosen yet. At first it was Debian, but Adrian has changed his mind.

One of the impressions I get from listening to the ogg streams is that the people at the workshop are very knowledgeable about the kernels out there and will choose the best kernel for Voyager. I agree with Adrian that writing a new kernel is not realistic and we desperately need to attract developers from other projects like KDE to contribute to writing the new WPS and SOM.

Finally, I am a user of OS/2 and at best a dabbler in programming. My contribution to OS/2's survival now or in the future is to buy OS/2 software and to contribute small amounts of money to the cause. It would be a sad day indeed if we lose the best parts of OS/2 if Voyager fails.

cytan


Subject  :  Re:Voyager and taking over the world :-)
Author  :  osw
Date  :  17 Apr, 2006 on 20:12

Thanks for your pessimism.

You are welcome



It's simply one possibility if you consider 32-bit operating systems....

Not really. I was not considering to replace superior 32-bit Os/2 with another 32-bit system... at least not with dos which was about to be replaced with os/2 "at the beginning" (remember? "better dos than dos...")

We don't need to replace 32-bits with another 32-bits. Do we? What for? Isn't os/2 already 32-bit?



That's really not the same. At first ReactOS isn't a simple "NT clone" but a Win9x compatible operating system.

It is "nt" (2000,xp,2003) clone, it is supposed to be compatible with nt api, and drivers - and supposed to run nt (2000,xp,....) apps...


OS/2 support could work through a subsystem, more or less equal to the Win9x part

Yes, it could of course - one would have to "just" write that subsystem (bit like a writing entire os....)


Why should it be impossible to support PM and 32-bit OS/2 applications (which of course should be the main focus)?? WPS is just an user interface - it can be added to every (gui) OS.

Ha!
I'm sure many of us believe it should be possible. But why to run it under "nt clone".



According to the ReactOS homepage there are already some people who considered writing a subsystem for OS/2. Does anyone know more about such efforts?

No, I don't. But I "consider" to become bilionaire



ReactOS probaply isn't based on MS code at all. They were just blamed for it. Currently they completely check their code to eliminate this rumor. This process is completed 70% as of now.

Ok, I'm not going to blame them also, i don't care and don't use windows nor consider to use reactOS.



It's already usable - not for the everyday work and not for the public but they got quite far yet. Just look at the screenshots section of their homepage

(http://www.reactos.org/xhtml/en/screenshots.html)


Well, I did look - pretty much the same impressive as ODIN-at-work screenshots... but still far from usability. But it's just my opinion.



You know that Voyager won't be just an OS/2 compatible WPS replacement but - as far as I know - a windowing system and WPS-clone for Linux/Unix. It probaply won't be OS/2 (backwards) compatible at all. It would be similar to GNOME as it was once planned but without X11.

No, I don't.



OS/2 compatibility is in my opinion a precondition for an "OS/2 clone". That's why I'm really not against osFree - it's a clean attempt. But building an operating system nearly from scratch is quite difficult and lengthy - maybe too difficult for our weak community.

I agree. But building os/2 subsystem for existing... let's say windows "NT" clone, or tailoring it to run on top of dos - musn't be easier nor shorter at all.



This was my personal opinion - and they were just examples/possibilities.
Robin

That's why I do respect them and polemise.

osw

greetings/2


Subject  :  Re:Voyager and taking over the world :-)
Author  :  Sebadoh sys3175@optonline.net
Date  :  18 Apr, 2006 on 01:44
The problem is that you take away usefullness now, to create something which if sucessful will be useful years from now. Just putting the WPS on another OS isn't exactly why I use OS/2, actually to be honest, I don't use OS/2 because of the WPS. The WPS is buggy, slow and quite possibly the worst part of OS/2. Yes, the configurablity is great, and I love the simple and clean applications but it is the biggest mess in the whole OS. I use OS/2 for stability (with except of the wps freezes), speed, and low system requirements. (of course my system is a Athlon 64 (754 venice core) 2 gigs of ram, matrox G400max, 2 dell 20in lcd monitors, and nearly 1 tb of hd space) but I like the fact that after boot nearly 99% of my memory is free after OS/2 boots. I don't mind the idea of voyager, but it has to bring binary support for OS/2 applications, a clean implementation of the wps, good device driver support, something similar to snap and uniaud and would be nice if it included network drivers in the same way, and everything GOOD in OS/2. Fast, reliable tcpip w/o all the exploits of linux or windows, samba would be nice, if it included a gui configuration similar to IBM's (of course without the freezes), the ability to use OS/2 or atleast windows print drivers (linux printing sucks). I mean.. it is so much, for so few developers which at this time would be much more useful fixing what is already here. Course, if something solid starts to come of voyager, I will be there working on the cause as well.


cytan (16 Apr, 2006 04:23):
From what I understand the kernel has not been chosen yet. At first it was Debian, but Adrian has changed his mind.

One of the impressions I get from listening to the ogg streams is that the people at the workshop are very knowledgeable about the kernels out there and will choose the best kernel for Voyager. I agree with Adrian that writing a new kernel is not realistic and we desperately need to attract developers from other projects like KDE to contribute to writing the new WPS and SOM.

Finally, I am a user of OS/2 and at best a dabbler in programming. My contribution to OS/2's survival now or in the future is to buy OS/2 software and to contribute small amounts of money to the cause. It would be a sad day indeed if we lose the best parts of OS/2 if Voyager fails.

cytan



Subject  :  Re:Voyager and taking over the world :-)
Author  :  Edu eduard@totcontes.com
Date  :  18 Apr, 2006 on 11:03
Well, in BSD world's, NetBSD has support for multiple binary executable formats ... like PE. Also, they've made a library which allows you to run Linux Applications under Free/Net/Open-BSD.

The main problem about BSD's are drivers. There are now more available drivers for OS2 than for other BSD's ( except for drivers included in UniAud and X11 ).

Or just look at AROS, Amiga Research OS ... what they have done ... they implemented first an API, and a Desktop Environment under a Linux system, and after getting it, they've started to do its own kernel, but knowing they can work with they desktop.


Subject  :  Re:Voyager and taking over the world :-)
Author  :  cytan
Date  :  18 Apr, 2006 on 15:31

Sebadoh (18 Apr, 2006 01:44):
The problem is that you take away usefullness now, to create something which if sucessful will be useful years from now. Just putting the WPS on another OS isn't exactly why I use OS/2, actually to be honest, I don't use OS/2 because of the WPS. The WPS is buggy, slow and quite possibly the worst part of OS/2. Yes, the configurablity is great, and I love the simple and clean applications but it is the biggest mess in the whole OS. I use OS/2 for stability (with except of the wps freezes), speed, and low system requirements. (of course my system is a Athlon 64 (754 venice core) 2 gigs of ram, matrox G400max, 2 dell 20in lcd monitors, and nearly 1 tb of hd space) but I like the fact that after boot nearly 99% of my memory is free after OS/2 boots. I don't mind the idea of voyager, but it has to bring binary support for OS/2 applications, a clean implementation of the wps, good device driver support, something similar to snap and uniaud and would be nice if it included network drivers in the same way, and everything GOOD in OS/2. Fast, reliable tcpip w/o all the exploits of linux or windows, samba would be nice, if it included a gui configuration similar to IBM's (of course without the freezes), the ability to use OS/2 or atleast windows print drivers (linux printing sucks). I mean.. it is so much, for so few developers which at this time would be much more useful fixing what is already here. Course, if something solid starts to come of voyager, I will be there working on the cause as well.



I guess there might be a fork in the road for you and you will have to make a choice in the future if Voyager ever comes into fruition. From the workshop ogg's again, there was a discussion about binary compatilibility with the current OS/2 and there are possibly two solutions: ODIN type solution or virtual PC type solution. Unfortunately, sticking with the current OS/2 is not a viable long term solution. At some point, BIOSs, CPUs will no longer be compatible with OS/2 (like the ISA bus disappearing) and we have no choice but to move. It would be nice if Voyager were there to catch us.

cytan


Subject  :  Re:Voyager and taking over the world :-)
Author  :  Fahrvenugen
Date  :  18 Apr, 2006 on 19:00
I admit I havn't had a chance to listen to the OGG, but I admit it has come to mind that an Odin type of solution might help with the binary compatibility issue. Especially if a method can be developed of "converting" EXE's and DLL's to a format that will run on the kernel which is chosen - similar to how Odin can convert Win32 EXE's and DLL's to something that can try to run under OS/2.

Of course, as with the current experience with Odin, there are still a lot of variables which take a part in determining whether or not such a converted app will work or not.

For me - in the short term - OS/2 continues to run on my system, and at least for the immediate future I'm still able to find hardware that it can run on in the event that I need to replace my desktop (which I don't see happening right away for a little while... things are running well for me). In the longer term, I can envision a day when I may need to run OS/2 in a virtual machine, after having migrated to something else.

As for IBM, I think we need to recognize that we're not going to get the OS/2 source (even stuff that IBM owns 100%) *unless* there are components of that source which could offer benefit to Linux. And in the case of source that could benefit Linux, its going to have to be IBM making the initiative - chosing to release that code (as they have done with JFS, Object REXX, and a few things like that... on that note I'd love to see them offer the SOM source code... at one point I think they even had a build of SOM that would run under AIX... it'd be great if they released that source to the OSS community... but if they did that, SCO would probably try and delay things even more by trying to add that to their lawsuit... after all SCO appears to be trying to claim that they own everything that ever ran on a *nix based platform... and even some ideas that never got put into code... but that's a different discussion... anyways...)

Admittedly - considering that the PPC version of OS/2 was based on a Mach kernel, it would be nice if we had that source... while it would take some time to port that to another platform, I have a feeling (based on my understanding of MACH stuff) there might be a better chance of getting the MACH version to run on another platform... But I don't see that happening any time soon either.

Just my thoughts


Subject  :  Re:Voyager and taking over the world :-)
Author  :  Sebadoh sys3175@optonline.net
Date  :  19 Apr, 2006 on 20:22
I am not sure what I would be moving to, but the more I look at voyager, and the discussions I have read about it, and the technology discussions in particular, I don't think I'd really want to have anything to do with it. Rewritting som? (why not just use som), using bits and pieces of over a dozen open source projects? Hoping it MIGHT all work together. It is actually not what I want to step into. And with all this they are really only talking about a window manager, not even an OS. So wait, we get to run linux applications with all the non-consistancies they currently have on an OS/2 like interface. The prettiest interface in the world cannot possibly fix linux.


cytan (18 Apr, 2006 15:31):

Sebadoh (18 Apr, 2006 01:44):
The problem is that you take away usefullness now, to create something which if sucessful will be useful years from now. Just putting the WPS on another OS isn't exactly why I use OS/2, actually to be honest, I don't use OS/2 because of the WPS. The WPS is buggy, slow and quite possibly the worst part of OS/2. Yes, the configurablity is great, and I love the simple and clean applications but it is the biggest mess in the whole OS. I use OS/2 for stability (with except of the wps freezes), speed, and low system requirements. (of course my system is a Athlon 64 (754 venice core) 2 gigs of ram, matrox G400max, 2 dell 20in lcd monitors, and nearly 1 tb of hd space) but I like the fact that after boot nearly 99% of my memory is free after OS/2 boots. I don't mind the idea of voyager, but it has to bring binary support for OS/2 applications, a clean implementation of the wps, good device driver support, something similar to snap and uniaud and would be nice if it included network drivers in the same way, and everything GOOD in OS/2. Fast, reliable tcpip w/o all the exploits of linux or windows, samba would be nice, if it included a gui configuration similar to IBM's (of course without the freezes), the ability to use OS/2 or atleast windows print drivers (linux printing sucks). I mean.. it is so much, for so few developers which at this time would be much more useful fixing what is already here. Course, if something solid starts to come of voyager, I will be there working on the cause as well.



I guess there might be a fork in the road for you and you will have to make a choice in the future if Voyager ever comes into fruition. From the workshop ogg's again, there was a discussion about binary compatilibility with the current OS/2 and there are possibly two solutions: ODIN type solution or virtual PC type solution. Unfortunately, sticking with the current OS/2 is not a viable long term solution. At some point, BIOSs, CPUs will no longer be compatible with OS/2 (like the ISA bus disappearing) and we have no choice but to move. It would be nice if Voyager were there to catch us.

cytan



Subject  :  Re:Voyager and taking over the world :-)
Author  :  phaelonimaire
Date  :  19 Apr, 2006 on 20:33
Making an operating system is a hard thing to do. Many projects like this fail (whatever happened to "Brix"?). However, my choice in operating systems was not made -just- by how my objects were oriented.
I like the CONFIG.SYS. I like the small footprint. I like the fact that is uniqe from the alternatives, and is far superior in many ways. Every OS/2 user has something inside them that makes them stay. This is the time for us to ask ourselves what that is, and where go from there. We will not progress into anything until we find out what it is we like, what we want, and what we're willing to do.
You may end up with two (or more?) camps of developers, OS, and WPS.
Don't let Linux's hardware portfolio blind you. Be very careful of what you can get for free. (Check the link below for more info)

We've lived a long time without complete hardware support. I would rather live with a limited hardware selection and create something that OS/2 reborn, rather than becomming another Linux distribution.

Operating Systems are hard. Here is a link to one such struggle:
http://www.upbeatlyrics.com/osideas/
(Not to mention some good (and unique) ideas)

Has someone started a discussion group for this yet?


Subject  :  Re:Voyager and taking over the world :-)
Author  :  lpino
Date  :  20 Apr, 2006 on 17:00
Well, I guess there are two questions that are relevant on this subject. What makes an Operating System? and Why we like OS/2?.
I was first introduced to OS/2 when some friends and I installed a beta copy of OS/2 2.1. Since the moment I saw that clock, I was hooked. In those days OS/2 was FAR superior to anything in the mass market. Windows 3.1 was a joke and DOS was really limited.
12 years later I continue to use OS/2, now at v 4.52 + Fx5. Funny but at the same time I started to use OS/2 I came across a copy of Linux, then at v. 0.93. Since then I have installed Linux on every computer I have had and everytime it doesn't last more than a couple of weeks installed. To me Linux ,SUCKS, it is unnecesary confusing. Come on! more than a decade and Apple in three years turn a UNIX into a beauty and Linux stills looks like a shame.
<rant off>
Well, to me what defines an OS is its API. Different API-> different OS. In the case of OS/2 its PM apis, Dos apis, Gpi apis,, etc.
When IBM ported OS/2 to PPC, they changed the kernel but kept the API. You have to remember that there are LOTS of software writen for OS/2 and that is a real plus. Binary compatibility it's a must (via emulation could be done too, but I rather recompile).
Think of all the little OSs that are trying to make it, like Syllable, SkyOS or even BeOS (Haiku), they lack software and rely on their POSIX subsystem to port as many programs as they can from Linux. OS/2 doesn't have that problem we have hundreds of native programs that we should use in a port.

Now to the question of what makes OS/2 special. Well by now the one thing that makes OS/2 unique is certainly WPS. Its something no other system has. I really like the multitasking but by now it's not that important, although it's still superior to Linux and Windows.
Why do I use it?. Just because I like it. I remember an article writen by a very known journalist at PCMagazine back in 1993 or 94, that was title "My wife thinks I'm nuts", in the article this person said that there was a secret message hidden on the screen that made him love OS/2. I guess I still see it, although he became blind or inmune to it, a long time ago.

Leonardo Pino


Subject  :  Re:Voyager and taking over the world :-)
Author  :  cytan
Date  :  20 Apr, 2006 on 18:14

lpino (20 Apr, 2006 17:00):

Now to the question of what makes OS/2 special. Well by now the one thing that makes OS/2 unique is certainly WPS. Its something no other system has. I really like the multitasking but by now it's not that important, although it's still superior to Linux and Windows.
Why do I use it?. Just because I like it. I remember an article writen by a very known journalist at PCMagazine back in 1993 or 94, that was title "My wife thinks I'm nuts", in the article this person said that there was a secret message hidden on the screen that made him love OS/2. I guess I still see it, although he became blind or inmune to it, a long time ago.

Leonardo Pino


I think you've hit the nail on the head. If the WPS is ported, if most of the current OS/2 apps can be either run under emulation or native I think we are there! If this path allows us to use the latest hardware and increases the user base, why do we care what the selected kernel is?

cytan


Subject  :  Re:Voyager and taking over the world :-)
Author  :  flywheel flywheel@worldonline.dk
Date  :  20 Apr, 2006 on 19:30
Neither OS/2 nor eCS has what you might decribe as a long term based future. I will keep eCS on my desktop (And I've planned to put it on my home server, when I have saved up enough to buy some new hardware). But the thought of going GNU or Mac scares me a little.

I personally welcome the Voyager project, and hope that many of you will do that also. It is good that they do not suffer from the dillution that everything have to be written from scratch but want's to reuse as much as possible.

32-bit OS/2 binary compatibility IMO is not immediately necessary, as long as the OS/2 APIs are supported making it possible to recompile the applications for Voyager.
But IMO - going for the next world (AMD64) would be prefeable, instead of just replacing one 32-bit system with another.


Subject  :  Re:Voyager and taking over the world :-)
Author  :  Fahrvenugen
Date  :  20 Apr, 2006 on 20:10
While I agree the API is important, the one thing which would be offered with binary compatibility is allowing us to keep using all those applications for which the source isn't public, or the source is no longer available. There are a lot of apps which I have installed on various OS/2 machines that I'd love to continue to use *but* if it becomes a matter of needing to recompile them, that's just not likely to happen. Examples include: TrueSpectra's Photo<Graphics, Dadaware's Embellish, AWE (yes, I still use AWE in a lot of web development), the older (but still quite useable, and sometimes will do stuff that the newer version won't do) StarOffice 5.1a, etc...

Subject  :  Re:Voyager and taking over the world :-)
Author  :  phaelonimaire
Date  :  20 Apr, 2006 on 21:07
..So Voyager is going to have a 64-bit reproductions of the current GPI and Win API Calls? Are you going to reproduce the bugs (intentionally) so your apps will recompile?
What functions are you lumping into "WPS", anyhow? You're going to end up recreating the whole GUI system if you plan on any OS/2 PM apps to recompile and work (which may be none if your API is 64-bit). Remember that there is a difference between PM apps and WPS apps.

Subject  :  Re:Voyager and taking over the world :-)
Author  :  flywheel flywheel@worldonline.dk
Date  :  21 Apr, 2006 on 16:10
How easy would it be to obtain OS/2 binary compatability via virtualisation (fx Xen port + installed OS/2) ?

And then scrap OS/2 32-bit compatability on Voyager itself, concentrating on new 64-bit API's

And would it present an acceptabel solution (Haven't used Xen, haven't the foggiest idea on how it actually works) ??


Subject  :  Re:Voyager and taking over the world :-)
Author  :  phaelonimaire
Date  :  21 Apr, 2006 on 19:31
The problem with binary compataibility is that to do it properly, one would also have to create a 32-bit API that is exacly identical to the IBM version we use now. I suppose something like the ODIN project could be developed to allow us to use the actual DLLs from OS/2, but that, too, would be difficult.
The root problem is that no one has made an effort to say what it is that this new version will be. Voyager is focusing on the WPS, however that is only a very small ammount of the puzzle.

Linux kernels can give you hardware support, but who's to say what file system will be used? Will there be EAs? Is everyone ready to recompile their kernel when changes need to be made?

Spend some time running Slackware Linux and tell me that this is what you want. Unless you just make a pretty window manager for (some) Linux, this is what you will have to look forward to. (Except it will be much worse)

Personally, I would be more interested in a entirely new operating system - something that is 64-bit only, and is designed to take advantage of new hardware. Some things can be borrowed from other places, but I think that the only true replacement for OS/2 is something that is truely as ground-breaking and robust as ..OS/2. I am not against the Voyager project - I just think that we need to realise that there is no way we're going to piece-meal an OS out of stuff laying around on the internet ... And it be worth anything. ..Let alone being called the true successor to OS/2.

Voyager in itself is a massive undertaking. When I said that there may need to be two (or more) camps of development, I meant that there may need to be a core OS group, and a GUI group. This would follow along the lines of OS/2 1.0.

You don't have to worry about hardware support, either. If the design is done properly, new hardware will be simple to add, and with Linux being open source, the handling routines can be easily added. It is very easy to make a device driver. ..It's hard to make an OS/2 device driver because of the lack of information and tools.
Remember: The hardest thing to work with is something that someone else made.

We're not going to come up with a solution tomorrow. Or next week. I don't even own a 64-bit machine. However, this kind of project, properly managed, will increase interest in OS/2, not decrease.

All I have remaining to say is this: IF somone is to make a new OS/2-like OS, let's at least attempt to do it right. I am ready to discuss and move forward. If things work out, even write code. My resources are available. All I need is a catalyst.


Subject  :  Re:Voyager and taking over the world :-)
Author  :  Sebadoh sys3175@optonline.net
Date  :  23 Apr, 2006 on 04:33
Do you happen to have an application which requires more then 2gb or ram? or is there some other reason you need 64 bit OS? heh.. I've got a sweet athlon64 setup right here, running fine on OS/2. (btw.. it runs better then xp64 speed wise.. though firefox is infinently faster in windows)


flywheel (20 Apr, 2006 19:33):
Neither OS/2 nor eCS has what you might decribe as a long term based future. I will keep eCS on my desktop (And I've planned to put it on my home server, when I have saved up enough to buy some new hardware). But the thought of going GNU or Mac scares me a little.

I personally welcome the Voyager project, and hope that many of you will do that also. It is good that they do not suffer from the dillution that everything have to be written from scratch but want's to reuse as much as possible.

32-bit OS/2 binary compatibility IMO is not immediately necessary, as long as the OS/2 APIs are supported making it possible to recompile the applications for Voyager.
But IMO - going for the next world (AMD64) would be prefeable, instead of just replacing one 32-bit system with another.



Subject  :  Re:Voyager and taking over the world :-)
Author  :  phaelonimaire
Date  :  23 Apr, 2006 on 15:21
Do I have programs which require more than 2GB of ram? Well, no - I use OS/2.

Subject  :  Re:Voyager and taking over the world :-)
Author  :  RobertM
Date  :  24 Apr, 2006 on 14:51
Very true... I dont have any true Warp (ie: not Win ported) programs that require such massive memory requirements, though I do have a few that use them (or that the capability to is appreciated) such as MySQL, numerous REXX scripts, audio and image editing software, etc. It is nice to know that the apps involved generally take a few megabytes of RAM allowing the rest to be used by them.


On another note, to the Voyager team... please work on HPFS/HPFS386 support even if it's as an IFS. JFS would be my second choice, but I still find the HPFS line more stable, and HPFS386 far faster and less quirky in many respects. Everyone else (including me) has already mentioned the thread and process handling in OS/2 that is second to none (out of currently available OS's), as well as it's multi-CPU scaling capabilites... so I wont mention that again.

- Robert


Subject  :  Re:Voyager and taking over the world :-)
Author  :  flywheel flywheel@worldonline.dk
Date  :  24 Apr, 2006 on 19:26

Sebadoh (23 Apr, 2006 04:33):
Do you happen to have an application which requires more then 2gb or ram? or is there some other reason you need 64 bit OS? heh.. I've got a sweet athlon64 setup right here, running fine on OS/2. (btw.. it runs better then xp64 speed wise.. though firefox is infinently faster in windows)

Well actually no - but is that the only reason to embrace AMD64 ?
We are talking about the portential birth of a new operating system, whoose binary OS/2 binary compatability most likely will emerge through a Xen solution anyway.
I say, if Voyager is going to be a fresh start - lets make it a real fresh start - leaving the struggle of implementing 486xs compatability in the trash can.
Also - what is going to stop it from being used in a server role ? Then wouldn't it be nice not to have to alter (AFAIR there is a hack that you can add to your programcode, that circumvend the 2Gb limit.) your applications in order to enable a more-than-2Gb-limit ?

Hell no - we won't port our applications to IA64 - Itanic runs IA32 just fine. (OK, a bad example - HAH!)

And no - actually everything runs on K6 & K7 processors in this house.


Powered by UltraBoard 2000 <www.ub2k.com>