OS2 World.Com Forum

Subject  :  What should be the next step for eCS? -> None of the above (2)
Author  :  cbo
Date  :  07 Feb, 2005 on 14:52
Ok, let's try another starting point, but still referring to the original thread of that title!

1. Developer's tools is indeed an issue! But it's NOT a question of deciding between GCC, VAC/C++, Watcom etc. It's rather the lack of "secondary tools" that limits development efforts!

- Tools for generating easy database access without coding everything "by hand"

- The same for GUI features

etc.

Both are existing - in the form of the "good old" VAC/C+ packages. But they are neither fully finished by IBM nor "officially" available any more!

The "bare bone" compiler can certainly be replaced by something like Watcom and GCC, but not the parts making the development of GUI or data access tools (to just name a few) really a pleasure. Of course there are things like wxWindows, but a half-finished Windows-oid GUI tool will never bring us the necessary tools to do really nice PM and WPS programming.

Conclusion (1): We would actually need VAC - finished and available, which is not possible! The second best thing to get is a good platform where developers can learn to make the best from the existing tools and "partial solutions". The place where this already happens and that can be supported is: Netlabs (http://www.netlabs.org).

2. There are still so many people talking about OS/2 in a "market issue". Not realizing that at this time a system like OS/2 for the price it actually costs has not the least chance at all to get any standing! NOBODY in the broader market pays any money for an operating system these days! I know people will debate this point, but still: Windows is normally "included", and Linux is free anyway, others follow. So there is not the least incentive for NEW customers to BUY something like OS/2 or eCS!

Of course I know that people at Serenity need something to live, and that IBM would probably not release OS/2 to the public that way. (Whether they "cannot", as many claim from time to time, or only "do not want to" is uninteresting, since the effect is the same...). But still: That's no argument for a young informatics student, interested and able to "try something new", but not ready or interesting to pay a few hunderd $ or EUR for it; he will look elsewhere.

The only way I could see how OS/2 or eCS could become again some kind of a "business case" would be: make it part of the one or other "service". While I cannot imagine any medium or large company to "switch to OS/2 or eCS" just because it's such a nice system, I could still imagine they could buy a system with eCS on it if it was part of a "service" they are buying from an external provider.

Conclusion (2): The necessary "next step for eCS" cannot be discussed "in the blue air", but it would have to be part of a specific "service" for which eCS would be offered at a given moment. This because the bare bone operating system is nothing to become a great success at the current time - and at the current price (fair or not)!


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  Sebadoh sys3175@optonline.net
Date  :  08 Feb, 2005 on 05:11
The entire point is that we need to migrate AWAY from IBM VAC, unless IBM was to do something like open source it which isn't something I see happening, but wouldn't be impossible.

If the tool doesn't exist, it can be created.

Toolkits need to be completed for GCC/OpenWatcom

Opensource/PD replacements for propritary tools must be found

a replacement for MASM for creating device drivers.

There is alot of work, and I can repeat that I never see OS/2 or eCS as a viable mainstream OS ever again (if it ever was). I love OS/2, been using it since 1989 when I bought my first PS/2 with OS/2, and have used every version since 1.1. The truth of the matter is we need to do the little things to make OS/2 stronger.

Finish projects to aid the porting of applications from both *nix, and Windows.

The idea of "OS/2 born and only OS/2 software" is something which can never come to fruitation especially as the user base deminishes. When it comes down to it, odin, everblue and the likes are native applications, even if they are not IBM's apis. Instead of denying the possiblities that these offer, we should embrace them, enhance them and improve them. Enhance the code, work on performance enhancements.

In other words, like it or not OS/2 will only exist without the aid of IBM. IBM is moving away from hardware, moving away from most of their software. They are aligning themselves as a services company, a patent licensing super power.


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  Sebadoh sys3175@optonline.net
Date  :  08 Feb, 2005 on 05:15
I use IBM OS/2 / eCS as part of a package that I offer, it is a point of sale, inventory management and accounting software. The cost of the base OS is part of the price of the system, which depending on customers existing hardware may or may not be bundled with hardware. It is one of the most inexpensive choices in the market, offering custom solutions for various clients of mine, currently working on using some old touch screen monitors which are compatable with OS/2 as an input option.

I believe you are right, if we want to get OS/2 / eCS into buisnesses, into the mainstream we need to integrate it for a particular task, sold with hardware, software. Everything you need to achieve a particular task.


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  Cornelis B.
Date  :  08 Feb, 2005 on 09:17
In *general* I would agree with you fully: OS/2 will only exist without the aid of IBM. And the faster everybody realizes that the better!

On the other side we still need everything we can get from IBM - because of lacking resources to do everything new from scratch! This is why I do NOT consider migrating my PmAs astronomy software project (see home.datacomm.ch/cbockem - at this site untouched for about 3 years now, but still "under construction" almost every day here...) "away from VAC", but instead I even now introduced the OCL, so binding it even more to VAC 3.08.

Why? With limited time ressources (family, job, other hobbies...) I would never be able to ever finish any unneccessary "migration" project! And why not use existing tools? VAC is abandonware by IBM, but it CAN be purchased the one or other way. And it can make the development of GUI interfaces much easiser!

This is no argument against other existing tools like GCC or OpenWatcom, which may be much more useful if it comes for example to porting existing software from Linux or Windows to OS/2.

Which brings me to your second point where I agree 100%: Odin and Everblue are for me the most important OS/2 projects (but not in the category "none of the above" of course!). Not because I do not like "native programs", but because I realize that priorities of many developers are not with the operating system, but with a certain task!

Again my astronomy example. I have for example the program "Guide Star Chart 8.0". A great astronomical mapping program with a huge lot of features. There is still an older DOS version, but the current version is developed for Windows only! And this not because the makers are "Billy fans", but because they simply take the most used platform in order to get things running for as many astronomers and hobby astronomers as possible! And for exactly that purpose they would never migrate the program to any "niche operating system" only because that is so nice or whatever. So in that specific case I am not very glad I can run the software with Odin very nicely! "Native" or not (which is a phantom discussion for me anyway).

The same with another software "XEphem", which is a bit "the standard" for astronomy on Linux platforms. But also this runs on my OS/2 computer - currently thanks to HobLink X11, but maybe once even better with Everblue?

So my point is: Even if we get a set of "mainstream standard applications" natively running on OS/2, we will NEVER also get all kinds of special software running - but this software is in many cases the "killing argument" for any OS/2 / eCS utilization!

So Odin and Everblue are not "betrayal", but they are really making OS/2 and eCS really viable OS alternatives!


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  Cornelis B.
Date  :  08 Feb, 2005 on 09:20
> So in that specific case I am not very glad I can run the software with Odin very nicely!

I don't know how that NOT slipped into the above sentence! I wanted to say "I am very glad..."


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  Kim
Date  :  08 Feb, 2005 on 09:43
We all agree to that we will never see OS2 to be open sourced by IBM and the main reason is usually that there are to many secondary patents that IBM is not the caretaker of and by that it's a dead duck.
Ok, let's turn it all around - what components is a must or would we like to see as open sourced when speaking of OS2? Just to give an example, the old ATM (Adobe Type Manager) would this be something that we would like to see open sourced and would a direct question to Adobe asking to release this old code? Might be that they say, well no problem, but then IBM has to agree as well, at least then one of the players is interested to open source an old code base.

Does any one have the full picture of above of what components that owned by IBM and not?


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  cyberspittle cyberspittle@yahoo.com
Date  :  08 Feb, 2005 on 20:02
CBO,

Thank you for taking the lead on this and helping Kim with this subject. It seems like everyone has an opinion they want to share and exploit every opportunity to put it forward.

We can't look back, unless we are trying to understand something that has occurred, or to explain it to someone who has jumped on the scene.

I think compilers is not an issue, and has never been (see my earlier post about the Warp 3 days). I personally prefer PM programming and C, rather than VAC and C++, due to my origins of learning to program with OS/2. We need secondary tools. MASM is not a viable option (look at all the proprietary code in OS/2 owned by Microsoft - your resource compiler show some trademarks, BTW). NASM looks good as it is current and up to date. We have Dr. Dialog for some stuff (search OS2VOICE for info).

We really need an IDE that is readily available by everyone, not just those with cash for VAC (hey ... I have a family with two kids to feed and I'm last on the list!). The GCC compiler for DOS (DJGPP) had an IDE called RHIDE that was modeled after the Borland Turbo C++ IDE. This has been ported to Linux (I'm not sure how popular it is/was). Perhaps we can do something similiar?

I had started on PM programming, using RHIDE as a prototype, in order to learn C programming. The Presentation Manager was the easiest way for me to learn OS/2 programming as there was/is an abundance of used OS/2 programming books on the internet and in used book stores. I would be willing to shell a program, but would need a graphical idea (or an example program in Windows/Linux to use as a prototype).

Here to Help (HTH)


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  aaronlawrence 0m3aces02@sneakemail.com
Date  :  09 Feb, 2005 on 16:52
I'm not sure I'd say VAC's GUI design is that great. The OCL is pretty good but basically you're still using dialog editor IIRC. You can use visual builder but that is a monster.

I like Sibyl, but that has the same problems: it is incomplete/buggy, unsupported and can't be officially bought. It also has the "OMG Pascal LOL" problem. On the other hand it is actually open source and is good for doing GUI designs.

Frankly, while I totally agree with the problem of GUI design being difficult, I don't see any appropriate solution that is within reach. I guess that the effort would best go to a very well known, not-too-complex open source thing like wxWindows.

Problem is, as toolkits get more advanced, they tend to start relying on features that are not available on less popular platforms; e.g. advanced C++ features, or other libraries like GTK. Then the "other platform" version gets out of date and becomes an orphan.

All this is just to say that OS/2 is in an extremely poor position.

Still, we'll all keep on, won't we


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  aaronlawrence 0m3aces02@sneakemail.com
Date  :  09 Feb, 2005 on 16:55


2. There are still so many people talking about OS/2 in a "market issue". Not realizing that at this time a system like OS/2 for the price it actually costs has not the least chance at all to get any standing!

I was going to suggest the academic edition of eCS, but then I see that is only $30 US off. WTF? The point of an academic edition is to make it a very low cost so that people can start using it early. THen again I guess SSI can't afford to lose money on a sale, and they no doubt have to pay IBM a fixed cost regardlesss.


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  Sebadoh sys3175@optonline.net
Date  :  09 Feb, 2005 on 19:16

Kim (08 Feb, 2005 09:43):
We all agree to that we will never see OS2 to be open sourced by IBM and the main reason is usually that there are to many secondary patents that IBM is not the caretaker of and by that it's a dead duck.
Ok, let's turn it all around - what components is a must or would we like to see as open sourced when speaking of OS2? Just to give an example, the old ATM (Adobe Type Manager) would this be something that we would like to see open sourced and would a direct question to Adobe asking to release this old code? Might be that they say, well no problem, but then IBM has to agree as well, at least then one of the players is interested to open source an old code base.

Does any one have the full picture of above of what components that owned by IBM and not?


I think if we could get the source to the ATM without the adobe code, that could be used to find a way to integrate a different truetype engine into it (freetype2). Without the hacks which Innotek has done to make it work (BTW, I am using freetype 2 as my ONLY renderer for truetype in my system, yes.. I read the warning and then proceded to add pmshell.exe to my list of truetype enabled applications).

I actually wasn't talking the viability of IBM open sourcing OS/2, I was talking about them open sourcing VAC and OCL. That would be a major step forward since OCL could then be ported to GGC or OpenWatcom, or the bugs could be worked out of the newer VAC releases, and give us yet another option for working with C++ in OS/2.

The same could go for MMOS2 and IBMLAN. We don't specificly need every piece of copy written code that they can't release, give us what we can, we can fill in the pieces. It would be alot better to extend and enhance what currently is embedded to to add yet another layer of the same thing unless our replacement is a total replacement so the original can be 100% removed. IBMLAN could be integrated with sambad. Would be great if samba for OS/2 could be managed the same way as IBM lanman shares (which is probably about as good as it gets in any OS/2, not that it doesn't need more options, but the interface when working is very useful).

OpenGL should atleast be released to SciTech so that if they wanted to get a team together (even outside of their own staff) it could be hooked to their drivers and support 3d, this would go a long way to making both games, and cad software working in OS/2 since really what holds back odin in this regard is the ability of the base OS.

There would be many small ways where this could be of assistance to us, such as the USB drivers (newest releases) trying to get updates for OS2ddpak evern though they discountinued this service, they should allow say netlabs to take it over and keep it up to date as they make improvements. How much longer are they going to find it worthwhile to offer these for money online, when few if any users bother? So someone needs to talk them into making it free for us to enhance, and they could benefit from the enhancements and offer the enhanced versions themselves. It is all a matter of how this is approached (and how much money the company representive that approachs them has. lol). But if we could get the source to LVM, we could fix the bug which rewrites the partiton tables on ipods making them still useless even though they can be mounted since the release of USB 2 MSD. (i've use the ipod as a removable harddrive and can use it to tranfer files, but the device itself complains of a corrupted partition table, and is unusable for music).


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  Sebadoh sys3175@optonline.net
Date  :  09 Feb, 2005 on 19:25

aaronlawrence (09 Feb, 2005 16:52):
I'm not sure I'd say VAC's GUI design is that great. The OCL is pretty good but basically you're still using dialog editor IIRC. You can use visual builder but that is a monster.

I like Sibyl, but that has the same problems: it is incomplete/buggy, unsupported and can't be officially bought. It also has the "OMG Pascal LOL" problem. On the other hand it is actually open source and is good for doing GUI designs.

Frankly, while I totally agree with the problem of GUI design being difficult, I don't see any appropriate solution that is within reach. I guess that the effort would best go to a very well known, not-too-complex open source thing like wxWindows.

Problem is, as toolkits get more advanced, they tend to start relying on features that are not available on less popular platforms; e.g. advanced C++ features, or other libraries like GTK. Then the "other platform" version gets out of date and becomes an orphan.

All this is just to say that OS/2 is in an extremely poor position.

Still, we'll all keep on, won't we


However, GTK could be ported with the newer GCC as well as the newest GCC being more or less completely up to date with C standards. This is why I said we need to pick 1 or 2 versions of C compilers and work to get tools for them which would aid in creation of applications. Instead of having IBM VAC (various releases), OpenWatcom, some of us still with Watcom 10.5 and others with GCC. try to work it so one, the most easily supported gets 90% of all the tools effort. That would keep us from falling behind in standards, and give people atleast the option of porting the newest apps.


Right now OS/2 is in a pretty good position as far as this goes, it's just a matter of not ever falling behind again. We need to get OS/2 developers involved in the compiler projects, so that code changes which break OS/2 support are worked out at the time they occur.. and don't send us back to the drawing board as far as support goes.


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  aaronlawrence 0m3aces02@sneakemail.com
Date  :  09 Feb, 2005 on 19:29
Get the code to the Adobe Type Manager without the Adobe code?!!! Um, there isn't any.

IMHO, the commercial/legal realities are:

1) Even if there is only a LITTLE bit of someone else's code, IBM can't release any of the associated code

2) Sorting out which modules CAN be released legally, takes a fair amount of time, including lawyers and negotiating with third parties. That means money. IBM therefore has no motivation to do it, unless someone else pays for it, and even then they would probably be loath to get into it for fear of getting mired in old arguments.

IMHO, all we can do is increasingly "deep" hacks until eventually whole chunks are replaced. I've done my part, replacing the whole help manager subsystem, and E

If SSI were to approach IBM to open source something, they would have to know that
a) it is REALLY important
b) it would get good results
c) they would not be going back again a month later for more
d) there was a reasonable chance of success
e) There is not a simpler alternative

So far none of that has really been true. If they want to spend money with IBM, its better gone on getting them to do selected bug fixes. (My perception).

I don't like it either, but the things I can think of where opening it up would really help (PM -> SIQ) are impossibly large and difficult to consider.


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  Kim
Date  :  09 Feb, 2005 on 19:44

aaronlawrence (09 Feb, 2005 16:55):I was going to suggest the academic edition of eCS, but then I see that is only $30 US off. WTF? The point of an academic edition is to make it a very low cost so that people can start using it early. THen again I guess SSI can't afford to lose money on a sale, and they no doubt have to pay IBM a fixed cost regardlesss.

Well, how can you loose money on sales that you would never do? I mean IBM Nordic (read Denmark, Finland, Norway and Sweden) had a simple policy, if the educations units called and asked for licenses they would either give away number of server licenses or highly cut off prices just to get out of the door. Most of the time this kind of customer is not expensive, as long they didn't use the software for commercial production.

If one needs support and help, you buy a full licens, if you just want to check out the product and get along with. Well, download the software for free. I would be happy to provide SSI with diskspace and bandwitdh for ISO images of eCS!


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  Cornelis B.
Date  :  11 Feb, 2005 on 12:04
A number of replies to a number of postings above:

1. OCL, other GUI class libraries and other tools

Of course one can discuss the pros and cons of the one or other such library back and forth! But since there are no absolute criteria that are valid for everybody this can never end. (For example do I use the OCL GUI classes "by hand" at the moment because I cannot understand why TODAY there are still programmers in the world writing dialogs that cannot be sized! But eventually I might even switch to "Visual Builder": It was a "monster" at it's time, but with current hardware it runs great! Or in other words: It came a couple of years early...).

But it's not only class libraries; it's also about other tools: IDE, debugger, what you want.

My "conclusion" from this thread is that I will go to the Netlabs site and start there a section about developer's tools where EVERYBODY can contribute it's tools, comment them, discuss them etc. I know there is the "DevCorner" at Netlabs, but that one is not updated for a long while already. But Adrian Gschwend agreed to insert such a "tools section" into the current EDM/2 "Wiki" section, where eventually also the DevCorner stuff could be moved (or at least linked to).

The site where this would happen is: www.edm2.com (which actually leads you to a Netlabs site).

I see that as the best contribution for the OS/2 developer's community at the moment, and it is feasable, using existing ressources. I would make it more publicly known once I did the first steps, so developers could really go and contribute.

2. The "open source OS/2" discussion: Whether IBM really cannot do it for copyright reasons or whether they simply do not want to (because it costs and brings them nothing!) is a useless discussion in my eyes: It simply won't happen. The best we will get is pieces - but they can already be extremely valuable!

3. The price debate: When I said that eCS must not cost anything at all in order to get a standing this is just talking from a "very high" standpoint, looking over the current world wide "OS landscape": There's actually nobody earning any money with selling an OS (as far as I know!) - except Microsoft to a certain degree. But that sells itself - driven by the fear of 99% of the computer users "not to be compatible". (Or in other words: If Windows XP was just "one system among others" there's a good chance that they wouldn't earn any money any more either...).


Subject  :  Re:What should be the next step for eCS? -> None of the above (2)
Author  :  cbo
Date  :  28 Apr, 2005 on 08:49
I started now the new Developer Tools Section mentioned above!

You find it at the new EDM/2 site in "Wiki" format, or directly via

http://www.edm2.com/index.php/Developer_Tools_Section


Powered by UltraBoard 2000 <www.ub2k.com>