OS2 World.Com Forum OS2 World.Com Online Discussion Forum. |
|
| Alternative to "open source OS/2" request |
|
|
Premium member in user
     posts: 277 since: 10 Aug, 2004 |
|
1. Alternative to "open source OS/2" request |
|
|
| Hi everyone, I was thinking the other day about the current petition to IBM to open source OS/2, and thinking of how it is unlikely (in my opinion) that it will see much success resulting in much of OS/2 being open sourced. However I was also thinking about some of the real reasons that there might be a need to access source code: That of fixing bugs in the current code without having to do an expensive and costly re-coding of portions of the OS, and potentially adding new features which would not be possible to add without access to the code. And my thought was, why not petition IBM to allow Serenity - through non-disclosure agreements - to have access to the source? My reasoning is simple. With the latest announcement it is crystal clear (if it wasn't already) that IBM doesn't plan to do much of anything with OS/2 through December 2006, other then provide the occasional device driver and fix. It even appears that there may be no more fixpaks, with Fixpak 5 being the "last" big update to OS/2. IBM appears to want to rid itself of the product, and this is not inconsistent with other anouncements over the past few years. Serenity on the other hand has an interest in seeing the OS/2 code base continue for a while, and has an interest in seeing bug fixes and new features. If IBM were to let Serenity access to the source code after signing an NDA with IBM this would (from what I understand) help to both give Serenity the code access to be able to try to fix certain issues, plus it would likely meet IBM's legal requirements regarding OS/2's source. Even if Serenity did very little to enhance the base code other then fix certain issues which have been around for a while - the OS/2 community would likely end up further ahead then it currently is. There is precedence for something like this. IBM has (from what I understand) - in the past - contracted with outside developers to work on OS/2 code. IBM has (again, from what I understand) required those outside developers to sign NDA's in relation to any OS/2 source code provided by IBM. Serenity Systems would simply be treated as an outside developer in relation to OS/2 source. Other software developers have also taken this type of approach with products - hire certain external developers to code specific parts of a product. It isn't unheard of in the industry. Serenity could potentially negotiate with IBM to be allowed to hire its own developers on a case by case basis. NDA's with both Serenity and IBM would also be required. I know this isn't the "open source" option that some people would like, but in my opinion it is more likely that IBM would agree to an NDA with Serenity then a straight open source of the code. Furthermore this type of agreement between IBM and Serenity might help with some of the issues that have led to the current "open source OS/2" petition. Any thoughts? |
| Date: 13 Jul, 2005 on 18:07 |
|
|
Premium member in staff
       posts: 2298 since: 12 Jan, 2001
 |
 |
2. Similar thoughts on Alternative to "open source OS/2" request |
|
|
I was thinking something similar to your 'alternative' idea. If IBM does not open source OS/2, it would be almost as good if Serenity Systems could have the source code so they can open a 'company' to develope the source code/drivers/etc futher. The 'employees' could be non-paid programmers (or paid vey little just for the records) who sign an agreement not to release the source code to non-'employees'. Serenity Systems already has a licence to use the kernel. Why not allow them to actually see the source code to improve it? This is a good idea even if IBM does open source OS/2 (I have been told that this is not likely). I will be sending this idea to IBM as an alternative to open sourcing OS/2 (via the feedback section of the download part of their web site http://www.ibm.com .
---
BigWarpGuy - - - - - - - - - - - - - - - - - - OS/2 Warp-ed/eComStation-ed to the very end. - - - - - - - - - - - - - - - - -
|
| Date: 13 Jul, 2005 on 19:47 |
|
|
Premium member in user
     posts: 139 since: 15 Apr, 2004
 |
|
3. Re:Alternative to "open source OS/2" request |
|
|
| I'd like to see this happen, but two obtacles jut out. First; IBM obviously, wants OS/2 to go away... though not right now. First they want to transfer their OS/2 business clients to another OS that IBM supports. So; why would IBM want to take steps that are clearly counterproductive to their long term goals? Second; Why would SSI do this? They do not have the resources to take on such an ambitious undertaking and; How would it be funded? That is, I can't see how SSI taking on the burden of the OS/2 source code could net them any additional income. I don't like to be a downer, but I do like to be realistic.
---
|
| Date: 13 Jul, 2005 on 23:37 |
|
|
Premium member in staff
       posts: 2298 since: 12 Jan, 2001
 |
 |
4. Re:Alternative to "open source OS/2" request |
|
|
| If IBM won't open source OS/2 and the SSI alternative (even with my ideas as stated above) is impractical, support of the OSfree project would be essential in the continued use of OS/2. OSFree would be the Warp version of ReactOS. How can one succeed if all ones thoughts are on failing?
---
BigWarpGuy - - - - - - - - - - - - - - - - - - OS/2 Warp-ed/eComStation-ed to the very end. - - - - - - - - - - - - - - - - -
|
| Date: 14 Jul, 2005 on 14:25 |
|
|
Normal member in user
     posts: 36 since: 15 Jul, 2005 |
|
5. Re:Alternative to "open source OS/2" request |
|
|
| Is it possible to create a found to pay IBM to develop the parts of the OS in a list?. |
| Date: 15 Jul, 2005 on 13:35 |
|
|
Premium member in staff
       posts: 2298 since: 12 Jan, 2001
 |
 |
6. Re:Alternative to "open source OS/2" request |
|
|
Would we be able to gain enough funds to make it worth while for IBM to consider?
---
BigWarpGuy - - - - - - - - - - - - - - - - - - OS/2 Warp-ed/eComStation-ed to the very end. - - - - - - - - - - - - - - - - -
|
| Date: 15 Jul, 2005 on 14:56 |
|
|
Premium member in user
     posts: 139 since: 15 Apr, 2004
 |
|
7. Re:Alternative to "open source OS/2" request |
|
|
| I like the idea of OSFree and in one incarnation or another, it's been around for years. However, there seems to be an inbility to determine which way it should proceed particulary with respect to the kernel. Personally, I'd love to see an OS/2 clone under GPL. However, I think that the shear size of the undertaking is staggering and frightens off many who could contribute... and that coupled with the kernel issue...
---
|
| Date: 15 Jul, 2005 on 16:14 |
|
|
Normal member in user
     posts: 1 since: 15 Jul, 2005 |
|
8. Re:Alternative to "open source OS/2" request |
|
|
last updated at 15 Jul, 2005 18:08 (2 times) The osFree group have choosen the kernel for the new OS/2 -- it will be the L4 microkernel. (See the last news at the http://www.osfree.org/ website). The osFree team now has a developer which has an experience in L4 programming. It's Samuel A.Falvo. See his page on OS/3 developing: http://www.falvotech.com/projects/os3.php . Now Samuel is writing an OS loader as an alternative for GRUB bootloader, which is traditionally used for loading L4 with initial servers. Also he's finished the 'format' utility. The new OS will use the LEAN filesystem as an example of the simple but superior to FAT filesystem, other than HPFS. It's very interesting perspective for OS/2 Community to get a successor for OS/2, which will support portability across many processors, SMP support "out of the box", capability to easy create distributed clusters, easy drivers reuse from linux, orthogonal persistense, running many OS personalities in paralell, fast Inter-process communication which allow to create systems as fast as traditional monolithic systems and also device drivers running in user space which can be killed and restarted as well as ordinary user applications! This technology is far superior (and faster) than Mach microkernel (on which are based MacOS X, OS/2 PowerPC, WinNT 3.50 and many others) and considering that, osFree team decided to no more use ReactOS as the base for OS/3 system. I installed ReactOS under VirtualPC and it is several times slower than WinXP and eats about 64 Mb of RAM. It is monster (as the WinNT is) and i don't know what perspectives would have an OS/2 successor based on it  And another question: Opensourcing OS/2 is good thing. And the hope dies the last. -- Our hope is that IBM will eventually open the sources, at least partially. I know two petitions to opensource OS/2, and signed both. One petition is on this site and is quickly getting the new voices. But 8000+ voices is not very much. I can suggest to join voices with another petition at http://www.petitiononline.com/OS24FREE/ Another suggestion is to ask help from linuxoids to get their voices. OS/2 is getting many software ported from Linux, and linux users probably wish to get some parts of OS/2 ported to linux. Their thread scheduler is worse than OS/2's and they don't have anything like the WPS. KDE and GNOME is worse. They are already have JFS. You can say that this is the deal with devil, but we can get them interested only if they get their profits. And we will get OS/2 opensourced instead. And the last suggestion: We can add the addition to the Petition that if IBM can not opensource OS/2 (intel) at all, we could ask IBM to release OS/2 PPC sources as they were developed solely by IBM and surely have no Micro$oft code in them. The OS/2 PPC sources could be ported to L4 as the GNU HURD, which was ported to L4 from Mach. (And very successfully, L4 version has good performance compared with the Mach version) WBR, valery
Long live OS/2! Good ideas don't die.
|
| Date: 15 Jul, 2005 on 17:55 |
|
|
Premium member in user
     posts: 139 since: 15 Apr, 2004
 |
 |
9. Re:Alternative to "open source OS/2" request |
|
|
All good news to be sure.  I'm not familar with the L4 microkernel, but if it does all you've said then it makes me happy.  It's also good to hear that they've moved beyond that issue. While I still subscribe to the OSFree mailing list I haven't been following it for a while. It's time to readup methinks.  With regard to the PowerPC version of OS/2 I think there was a discussion on the newsgroups a while back about opensourcing it. Since IBM clearly has no intention of using it, it seems that it would be a FAR more likely target for opensourcing that the existing OS/2 for the X86 CPU. I know OS/2 for the PowerPC was far from complete and now, technically speaking, is far behind current versions, (eCS 1.2 for example), it still has to be a better foundation to build a clone on than starting completely from scratch. At the very least much could be integrated into the existing OSFree platform... but then again... I'm not a programmer.
---
|
| Date: 16 Jul, 2005 on 02:01 |
|
|
Premium member in user
     posts: 252 since: 09 Dec, 2002
 |
|
10. Re:Alternative to "open source OS/2" request |
|
|
| I think it is clear the eCS is the route to follow. If we can get enough people to upgrade or, at the very least, buy a licensed version of eCS this would infuse money into Serenity that could pay for projects in support of OS/2. I can understand people wanting to continue to use OS/2 Warp (v3 or v4), however if all OS/2 users have at least one licensed copy of eCS 1.2, we would all benefit. If you know a computer geek in the family that has x86 hardware, give them a licensed copy for their birthday or Christmas. If you are well off financially perhaps a bulk purchase and donation to a computer club or computer science department would be another idea. We also need to grow the numbers of developers to use the tools we have (open Watcom, GNU C/C++, etc.). Increasing the number of developers would enable us in the long term. Licensed copies is a short-term financial plan. As Steve Balmer once said in a now-famous internet video (while slobering on himself), "Developers! Developers! Developers!". This is how we can sustain ourselves. Linux has grown (not by GPL, although it has helped in accessing code) through the number of software developers. We need more software talent. |
| Date: 16 Jul, 2005 on 18:25 |
|
|
Normal member in user
     posts: 36 since: 15 Jul, 2005 |
|
11. Re:Alternative to "open source OS/2" request |
|
|
Would we be able to gain enough funds to make it worth while for IBM to consider?
I don't know, but we can make a petition form to know it. Particulars can contribute with little, but enterprises (like Serenity), with a great contribution. But before, there is a question: How much costs a year of development by IBM?. |
| Date: 18 Jul, 2005 on 10:06 |
|
|
Premium member in staff
       posts: 2298 since: 12 Jan, 2001
 |
|
12. Re:Alternative to "open source OS/2" request |
|
|
[quote]cyberspittle (16 Jul, 2005 18:25): I think it is clear the eCS is the route to follow. If IBM discontinues developement of OS/2, what future would eCS has without it? Since they do not see the source code (IIRC) that IBM developes, on what would they base future versions of eCS?
---
BigWarpGuy - - - - - - - - - - - - - - - - - - OS/2 Warp-ed/eComStation-ed to the very end. - - - - - - - - - - - - - - - - -
|
| Date: 18 Jul, 2005 on 14:13 |
|
|
Premium member in user
     posts: 277 since: 10 Aug, 2004 |
|
13. Re:Alternative to "open source OS/2" request |
|
|
BigWarpGuy (18 Jul, 2005 14:13): If IBM discontinues developement of OS/2, what future would eCS has without it? Since they do not see the source code (IIRC) that IBM developes, on what would they base future versions of eCS?
And that is part of my thinking to try and suggest that SSI have access to source code, through an NDA. It would give SSI the opportunity to both fix current problems and potentially further develop, while also potentially avoiding legal hassles for IBM of open source related to OS/2. Fortunately OS/2 is a very modular OS - and can be added to easily. But the difficulty that I see comes with trying to fix issues that exist (for example, some of the bugs in MMPM which have been around since Warp 3) and getting the system to work on newer hardware where it can't be made to work with new device drivers. Within discussions about IBM's withdrawl notice, I've seen some comments on various forums suggesting that eCS will now have difficulty, asking the (valid) question of where will SSI get new device drivers? But let's consider this for a moment. I can't speak for everyone, but when I think about my own system, there are a whole bunch of non-IBM device drivers running it right now. For example the non-IBM device drivers that I run right now: -SNAP for video -Uniaudio for sound (although this was initially developed from IBM code, continued development has come from others) -Danis506 for IDE -SIO2K for UART / Com port -CW-USB for USB -FAT32.IFS -DANIATAPI.FLT -DANIDASD and ASPIROUT -PCIND.OS2 (network card driver) -WCAST.SYS (for the Hauppauge card) -AMOUSE And on my laptop, the driver that runs the PCMCIA slot, along with the driver for the network card were both developed outside IBM. Yup, that's right. Even something as simple as a mouse driver, I find the 3rd party one works better then the one supplied by IBM. That's not to say that the IBM drivers don't work (or can't be made to work), it is just that I've found that for my hardware, 3rd party drivers seem to work better then the ones offered by IBM. To me it is obvious that people outside of IBM have figured out how to write device drivers and file system drivers. Where I see the initial benefit of convincing IBM to allow SSI access to OS/2 source - potentially through an NDA - would be the ability that SSI would then gain to be able to 1. Fix problems that currently exist (the previously mentioned MMPM is one area that could use some work), and get the OS to work on newer hardware (for example, the recent work which has been done on the kernel to allow it to work on 64 bit AMD processors). As for adding "new" features to OS/2, that ability already exists without the need to access source code. The object oriented nature of the WPS and other parts of the OS already allow for this. Just to quickly give my thoughts on the question of why IBM would do this when they want to keep their customers and move them to Linux, I think there are a few things to consider (and recognize, I do not speak for IBM nor SSI here). First, from what I understand there are 2 types of OS/2 customers. The smaller, non-enterprise customers - the end users, the SOHO, etc. For these customers, I don't believe that IBM cares one way or the other if they stick with IBM. Let's face it. For me, the money I might spend on a Warp 4 license as an individual is really insignificant in the overall IBM scheme of things. IBM's CEO makes more in a single hour then a single Warp license brings in. Its that simple. For customers like myself, IBM has been saying since they stopped marketing Warp 3 to home users (remember the Nuns?) that OS/2 really isn't meant for the home or smaller user. On the other hand, there are big "enterprise" customers. Companies which have 1,000 or more licenses. Often they'll also have service contracts which mean big $$$ for IBM. These are the clients that IBM doesn't want to lose. But on the other hand, these clients aren't likely to completely drop IBM in favour of SSI. Part of what these clients are paying for is a service contract. They may also have other (more expensive) IBM servers and even potentially mainframes. These companies work with IBM to ensure they have a system that *works* the way it is supposed to. Money isn't the first concern. On the one hand it is true that some of these clients may purchase eCS if - in the future - eCS is able to support hardware that OS/2 is no longer able to support - and these clients want to stick with their OS/2 infastructure. However I doubt this would be a direct decision for those clients suddenly "drop" their IBM contracts. On the other hand, I can also see situations where such clients work with IBM to switch systems over to Linux. It is for these reasons I really can't see IBM "losing" many enterprise customers if they allowed SSI access to the source. On the other hand, if the agreement were such that SSI allowed IBM to use any of the "fixed" code for its own clients - then IBM clients who have stuck it out with OS/2 (regardless of the fact that IBM withdrew the product) would be able to benefit from these fixes on a case by case basis. As to why SSI might do this, to me the reason is simple. It would allow them to have the ability to try and fix long standing issues which exist in OS/2 code. Let's be honest here. Most of OS/2's code is rock-solid and stable. Does what it is supposed to do. However there are areas which present difficulties which - unfixed - need to be worked around. Two areas that I've read have problems are the multi-media sub-system, and GRADD. SSI has done work to try and fix some of the problems with MMPM. Using a different set of codecs, getting a new installer developed. Using newer drivers such as UniAudio and the generic WinOS/2 sound drivers. However wouldn't it be nice if they had the option to hire someone to try and see if some of the other problems in MMPM could be fixed? True, they could likely hire IBM to do this too. However with IBM's announcement, and the fact that MMPM hasn't had any major fixes or updates since Warp 3 - I doubt IBM will be highly interested. As for GRADD, wouldn't it be nice if Sci-Tech had access to GRADD code to try and fix some of the stuff which I've heard is a "GRADD" problem and not a SNAP problem? Of course, an NDA would be needed for this too (assuming OS/2 isn't open sourced) As for costs to SSI, I don't see *major* updates happening to OS/2 code even if they had access. My guess would be we *might* see bug fixes, minor updates, maybe some new features. But then to be honest, if the code were open sourced, this is about the same thing that I'd expect. Simply because the OS/2 community (in my opinion) doesn't have the resources or numbers needed for "big" additions or changes. Would SSI be even interested in access to the source through an NDA? I really can't answer that question. That's best left for SSI to answer. So I won't speculate on that further. Anyways, just my $0.02. Again, I don't speak for either IBM nor SSI, so these are just my own thoughts. Feel free to comment! |
| Date: 18 Jul, 2005 on 18:34 |
|
|
Normal member in user
     posts: 1 since: 20 Jul, 2005 |
|
14. Re:Alternative to "open source OS/2" request |
|
|
aasdelat (15 Jul, 2005 13:35): Is it possible to create a found to pay IBM to develop the parts of the OS in a list?.
SSI is exactly the source of funding to pay IBM to develop parts of the OS. Or more specifically - to support parts of the OS. Remember - OS/2 is designed to be modified, and for pieces to be replaced. The WPS can be modified (xWorkplace shell), device drivers can be built by anyone, the TCP/IP stack can be replaced, Rexx can be replaced, new file systems can be installed (IFS - netdrive, etc.), MMOS2 can be replaced or modified. While there are some things that are clearly under the control of IBM - such as the kernal, there are all sorts of other stuff that can be modified and replaced. While Open Source may fix some issues on OS/2, it also has the very real danger of introducing issues that completely change the OS for the worse. A single example is backward compatibility. Each new release of OS/2 (and there have been new releases of OS/2 although they were released as fixpacks) runs software that ran on previous version of OS/2. I have found very very few software packages released for earlier versions of OS/2 that do not work on eCS v 1.2. There are not many operating system that can say that. This happens because IBM forces it to happen. With Open Source, no one can force anything to happen. The other problem with Open Source is: just because the source is open doesn't mean that anyone will work on it. I would much rather pay SSI $150 every year or so for the latest version of eCS, and have SSI handle the interface to IBM and the purchase of extended support from IBM for the portions of OS/2 that need it. I would rather the OS/2 community directed its resources to building and enhancing the applications that OS/2 needs (which IBM will never do), and let IBM do the support that they are willing to do for money. This will result in a much superiour over all OS/2 package then spending finite OS/2 developer resources on problems that currently (and in the future) can be solved by throwing money at IBM - through Serenity Systems. |
| Date: 20 Jul, 2005 on 01:53 |
|
|
Premium member in user
     posts: 252 since: 09 Dec, 2002
 |
|
15. Re:Alternative to "open source OS/2" request |
|
|
BigWarpGuy (18 Jul, 2005 14:13): [quote]cyberspittle (16 Jul, 2005 18:25): I think it is clear the eCS is the route to follow. If IBM discontinues developement of OS/2, what future would eCS has without it? Since they do not see the source code (IIRC) that IBM developes, on what would they base future versions of eCS?
Hey BWG, I think it is pretty clear that IBM will continue to make make mods to the OS to large customers with a sizeable bank roll. With SSI as the front for the home user, we stand a chance of bank rolling improvements. IBM will stop selling OS/2 Warp (2005) and customer service a year later (2006). Beyond 2006 it will only cater to existing customers; ie SSI (eCS). That is how we get what we want. |
| Date: 20 Jul, 2005 on 04:53 |
|
|
Normal member in user
     posts: 36 since: 15 Jul, 2005 |
|
16. Re:Alternative to "open source OS/2" request |
|
|
Ok, I understand that everything for what could be necessary a fund, is yet done between SSI and IBM. And that this will not change in the furture (it will not change beyond 2006). And the solution is to suscribe for a year of updates in eCs. This will give SSI the necessary funds to ask IBM to continue with the developing of those parts f the system that only can (and I also think like you: only could) be developed by IBM.And this idea can be ported to the developers of software for eCs. Ok. I will tell this to the guis at os2spain. |
| Date: 21 Jul, 2005 on 12:37 |
|
|
Premium member in staff
       posts: 2298 since: 12 Jan, 2001
 |
 |
17. Re:Alternative to "open source OS/2" request |
|
|
It would be worth paying for that kind os subscription knowing it will help OS/2 and eCS users in the future.
---
BigWarpGuy - - - - - - - - - - - - - - - - - - OS/2 Warp-ed/eComStation-ed to the very end. - - - - - - - - - - - - - - - - -
|
| Date: 21 Jul, 2005 on 14:24 |
|
|
Premium member in user
     posts: 139 since: 15 Apr, 2004
 |
|
18. Re:Alternative to "open source OS/2" request |
|
|
BigWarpGuy (21 Jul, 2005 14:24): It would be worth paying for that kind os subscription knowing it will help OS/2 and eCS users in the future.
But the question is; How many users will do that? and its followup; And for how long? Before you can answer that we have to answer this: How many OS/2ers are there right now? Few seem to know the answer to that and those that do aren't telling us.I have a significant, longterm financial investment in OS/2, (since 2.1), and I'll do what I can to support it myself. I'm sure some users will do the same, but are we enough in number to actually influence anything? I'd like to think so, but is that realistic? OS/2ers are notorious for being vocal when it's time for talk, but being tight-fisted when it's time to pony-up the cash. 8( I'd really like to see something breathe new life into this much loved and much maligned OS, (for the long term that is), and right now, all our hopes are in the hands of SSI.SSI is a great company that has truly extended the life of our darling OS, but are they big enough to carry us for the long haul? Keep in mind that their size and longevity is entirely dependant on the OS/2 user base. The only other thing I see for the long term is OSFree, may it blossom and flurish. 
---
|
| Date: 22 Jul, 2005 on 14:44 |
|
|
Normal member in user
     posts: 36 since: 15 Jul, 2005 |
|
19. Re:Alternative to "open source OS/2" request |
|
|
What does SSI say about eComStation beyond 2006?. What does IBM say about kernel and driver development for SSI beyond 2006?. Does anybody know this two questions?. Perhaps a SSi and a IBM worker?. |
| Date: 25 Jul, 2005 on 09:45 |
|
|
Normal member in user
     posts: 44 since: 05 Apr, 2005 |
|
20. Re:Alternative to "open source OS/2" request |
|
|
| Hi! All your ideas seems to be interesting... to spend bucks on persuading IBM to take some care of some parts of os for some large customers... Well my question is: why? Will SSi bring new life into aged kernel? I don't think so. Will ecs ever has 64-bit kernel? Again - I don't think so. So, what can we expect to have is little polishing gui, few drivers and even working acpi.... just imagine! I'm not against SSI as they doing good work, but they're rather preserving existing internals to run on modern hardware, but not with modern features. Ecs run on amd64 as ... a 32-bit os. And what about multicore cpus? What if they won't be smp compliant? Will they ever work for os/2? Again - I don't think so.... It's nice to use legacy system on legacy hardware to run legacy apps.... but if so, we already past... The only hope, to move forward in terms of compatibility with hardware and new features is to get os to new stage. New kernel, new api and of course new WPS. Why don't we talking about supporting those guys from osFree ... they've quietly choosed new kernel (l4) as a base for their project, which seems very, very promising as a new engine for os clone (yes it's more modern than mach base of macOSX or winNT). And it's architecture allows to use "virtual device drivers" so it would resolve worst warp pain... device drivers. We are talking much about flexibility and modularity of warp, so why not use it for totally replace all ibm or ms legacy. Step by step. But we need base. Kernel. So my two cents instead of $$$ for IBM or SSI. greetings/2 |
| Date: 25 Jul, 2005 on 10:51 |
|
|
| Alternative to "open source OS/2" request |
|
|
| All times are CET+1. |
< Prev. | P. 1 2 | Next > |
|
|
|