Public Discussions > General Discussion

What is the agenda for OS/2 today?

<< < (3/4) > >>

Doug Bissett:

--- Quote ---I don't see Mensys doing a job to open source OS/2, I don't see it on Roderick conversations and I don't see it the eComStation web site.
--- End quote ---

Well, I do see it happening. Do a Google search on "eComStation open source svn", and you will find a list of what is available, as OPEN SOURCE. Check those packages, and they will tell you where to find the source, including for ACPI and MultiMac. What you don't find there, are the binaries. They are in Software Subscription, until they are released in a new version of eCS. So far, everything, that I know about, has been included. There are also a few packages that Mensys has contracted to include in eCS (DragText), that are NOT there, but they are available to everybody on the shareware model. I also think that Mensys has been collecting source code for those packages, but will not make that available, because of a contract with the developer, that includes a way to open source it, under certain conditions.

Mensys has NOT been silly enough to try to rewrite OS/2. After all, it took IBM, with a large team, many years to get OS/2 to the state it was in when they quit, and they had the source code. What Mensys has been doing, makes a LOT more sense than what you propose. If eCS lasts long enough, and finances permit, Mensys will do more, and that may include rewriting parts of the OS. Meanwhile, they are being realistic about the whole thing, while you are not.


--- Quote ---Why do you do something about Multimac?
--- End quote ---

Actually, I have been doing something about MultiMac. I have been attempting to keep Genmac going, while we wait for MultiMac. Unfortunately, it seems that Genmac has come to an end (what works, still works, but newer stuff does not). I have also been testing new releases of MultiMac, and doing what is necessary to help the current developer(s) to fix the problems, that I find, although supplying them with hardware seems to be most effective at getting it to work (Mensys has been doing that). What MultiMac needs, is a developer to buy a new device that doesn't work with the current support. Perhaps they would be interested enough to go to the source, and figure it out. If we expect users to fund it, we might as well tell Mensys to do the funding model, and have them collect the money. Of course, that would mean that they would need to sell the final product, rather than supply it for free (like they did with OpenOffice), simply because those who did fund it would feel cheated if others got it for nothing. I think that is the main reason why it is so difficult to get funding for projects like Firefox.


--- Quote ---I want to set a goal for the WPS development but I haven't find a developers that wants to help on this yet.
--- End quote ---

I wish you luck. Perhaps it would be easier, if you would start with a less ambitious project, that some developer might actually know something about. Most, if not all, developers know that what you propose is going to tie up, at least 5, programers, for many years. Most of the first year will be to figure out what they are actually trying to do. Until they figure that out, it would be impossible to quote any numbers, like how much, or how long. So, you are likely looking a half a million dollars, just to get some realistic numbers to do the project. No developer will put that into a project for an OS that can't even do accelerated video, or WiFi. Users are not likely to fund it, because there is a good chance that the numbers will come back at an impossible amount, and there will be nothing to show for what has been spent already. Users are reluctant enough to fund Firefox development, where some progress has been demonstrated.

The only hope that you have, would be to find retired programers who actually worked on the project for IBM, and are looking for something to do. There should be lots of them who worked on small parts of the WPS, but I suspect that those who did the initial design are not going to be interested, and you would need a few of them, just to map out what it does (if anybody can remember all, or most, of it).


--- Quote ---but we are community of users, that's why we need to look to remain with the platform in time.
--- End quote ---

I hate to say it, but you are dreaming in technicolor. There are too few users to support anything more than what is being done today, and Mensys is funding a lot of what is being done (volunteers do the rest). It is also obvious, that Linux has caught up to windows, and will soon pass OS/2. Some say that it has happened already, and it has, in many ways. The WPS, on it's own, is not going to save eCS. More important would be to get the Kernel updated to 64 bit code. I would assume that that would also cure the memory space restrictions. Of course, doing that would mean that most other parts of eCS would also need to be updated. Drivers, for instance, are still 16 bit code, and, from what I hear, 16 bit code simply will not run on a 64 bit system. Prognosis: Not likely to happen, but that would probably be a requirement to make a new WPS actually worth the time, and effort. Patching up what we already have (the SMP patches), and adding parts (ACPI, MultiMac) that are essential to being able to actually use OS/2, is the only way to prolong the life of the product (short of the lottery scenario). You, Mensys, or even Microsoft, will never be able to stop progress in Linux,  and once the users get to the point where eCS can no longer do the job, they will be forced to switch to something else. Even if you had a new WPS today, it would only delay the inevitable by a short period of time. It is equivalent to putting a new convertible top on a rusty old Nash Rambler, with bald tires. It would make more sense to replace the tires, and get more life out of it, but doing either one, would not increase the value of the car, and it is only a matter of time, until it stops working.

Martin Iturbide:
Hi Doug.

You have your opinion, but I don't share that.  Do I have to reply to your post? or it will not make any difference on your way to see it?

The same thing "Why do you do something about the 64bits kernel? Why do you start finding developers and try to set and raise a money goal for that project, maybe something similar to what Silvan is doing. If it is going to be an open source project we are going to support it. "

I will focus on keep updating the EDM/2 Wiki, the OS2World Wiki, trying to organize more open source code at github.com/os2world and sharing the news at OS2World.com. You are free to help like everybody here.

Regards

Doug Bissett:

--- Quote ---You have your opinion, but I don't share that.  Do I have to reply to your post? or it will not make any difference on your way to see it?
--- End quote ---

How can I change my view? My view is supported by facts. Your view is  only a dream. I would like to see your view come true, BUT, I do NOT want to see you destroy what is left of OS/2, by sucking the life out of what is left of it. You will not find a developer who is willing to take on the job, for the funds that you will be able to attract (unless somebody wins the lottery, and donates the funds to the project). All of that effort could be put to much better use.

IF eCS can ever catch up to the basic expectations of a new user, we might get enough momentum to take on larger projects. Meanwhile, forget the 64 bit kernel, and the WPS, we need WiFi, Bluetooth, a complete USB stack (including USB 3.0), accelerated video, an up to date browser, and some other things. All of them are possible, and some are in the works, at Mensys, but it is difficult to find developers who can do the job, never mind those who will work for small compensation, simply because they also want to see eCS succeed.


--- Quote ---The same thing "Why do you do something about the 64bits kernel?
--- End quote ---

Simply because, the kernel would be an even bigger job than the WPS. Do YOU know anybody who has the skills to build a new kernel, that will do what the OS/2 kernel does?  I know that I don't. Therefore, we are back to borrowing software from Linux. If we do that, we might as well use Linux. There is the Russian group who are messing around with the kernel, but it has been a LOT of years, with little result. Then, there is the problem of all of the other things that would need to be updated to be able to actually use a new kernel. The full IBM development team would be required to get the job done in our lifetimes (especially if they really don't have the source). Despite what others seem to think, I still believe that the whole desktop platform (including laptops, as we know them), will be finished in the next few years, and it is very unlikely that OS/2 will survive as an OS that runs in a real machine. If it runs in a virtual machine, there would be no need for a lot of the ported software, because it would run better in the host machine.


--- Quote ---I will focus on keep updating the EDM/2 Wiki, the OS2World Wiki, trying to organize more open source code at github.com/os2world and sharing the news at OS2World.com. You are free to help like everybody here.
--- End quote ---

This is all good. It helps those who are doing, or learning, development, and it doesn't take away anything from what is being done, to extend the life of OS/2, today. Some of it may even be new enough, that some people might be able to learn a bit about programing from it.

Alex Taylor:
If I can interject my own thoughts.

I wholeheartedly agree with Martin's long-term objectives, and even the broader aspects of his strategy.  I do think it's unwise to try and start with the WPS, however, for several reasons.

First of all, the only way this project can get off the ground is if it attracts developer enthusiasm from the OSS community outside OS/2. That means, practically speaking, that the project needs to start with something that can actually be tangible and useful to the wider OSS community.  A new kernel is useful because it has no dependencies on lower-level OS/2 software.  Non-OS/2 users can actually do things with it.  Even a replacement for PM is potentially useful because, being itself the lowest-level graphics environment, it only depends on the underlying OS services – meaning that if it's well (i.e. portably) written, it could in theory be built and used on other OSes like *nix.  Again, it's something tangible for the OSS community to get behind.  But a replacement WPS that starts out by depending on the existing Presentation Manager... if you go and ask a random open source developer if they'd be willing to help out, they're just likely to respond "As a non-OS/2 user, exactly how do I benefit?"

In other words, what pitch is more likely to attract developers right off the bat?
1. "Let's create a new, open operating system based on the architecture of OS/2"
or
2. "Let's write some OS/2 software that only runs on existing OS/2 systems"

Second, and related, is that the necessary toolchain is so obscure that even if we can attract these outside developers, we'll have real trouble equipping them and keeping them engaged.

Third, and this is mainly a philosophical argument that I freely acknowledge many people will disagree with, is that I've long been opposed to the idea that the WPS is "the one indispensible feature" of OS/2. OS/2 is so much more than that, and I'm afraid that focusing straight away on the WPS above all else risks obscuring all the other parts of the operating system as equally important objectives.

Martin Iturbide:
My suggested plan is based that on that the projects to clone OS/2 that had tried to start "down to top" had lost traction, and none of the generated components are being used by the OS/2 community and are not included on eComStation.

While the open source projects that had been developed to run over OS/2 like XWorkplace and WarpIn had been included on the platform and seems to have more recognition on the community. But I also want to focus on replacing the close source components that are working on OS/2.

Sometimes when starting a development "down to top" the developer gets too tempted to fix what is wrong in OS/2, and to fix that it generated components that incompatible with OS/2. If we generate today a PM replacement that will turn incompatible with IBM's SOM and IBM's WPS will it be included on eComStation? Will it be used for the community?

Even that I want PM, SOM, WPS open sourced (cloned) in the long term (technicolor dreaming), starting with the three at the same time will not be possible - just see all the flame just by suggesting to start with one :)  I can agree and support to start with PM, but we need to make it sure that it will 100% compatible to what we have today and make IBM's SOM and WPS to run over it.

For example, if we start with a PM clone, I think that "FreePM" for "OSfree" is a good starting point. But sadly I can not asset how far away or close are from the goal.  I noticed that FreePM/OSFree generated some DLL replacements that worked on OS/2 but it was not a big part of PM yet.

Navigation

[0] Message Index

[#] Next page

[*] Previous page