---

March 17, 2005 - Wolfgang Draxler had the following comment regarding the March Editorial and a question:

First of all, thank you very much for the issue. There are many interesting articles again.

I also wanted to add a bit to the article "Developing for OS/2 - starting is easier than you think":

You wrote that you like Pascal better. There is the development environment "WDSibyl" that is available from myself and which is based on Pascal. It is a successor of Sibyl (Speedsoft). You can find it at: http://members.chello.at/webstar/wdsibyl.html

You also mentioned problems creating the INF file. Here my question is: Is there any program available for converting IPF to HTML files?

Christian Hennecke's answer:

No, not directly. One can first create an INF file and then convert and split this into HTML pages with Ulrich Möller's INF2HTML. Furthermore, there are applications like Vyperhelp which are capable of writing both formats. However, this is not an option for us. First, IPF has certain limits that HTML does not and which we do not want to impose on the HTML version. And second, submitted articles would have to be converted which would probably very complicated. For our purposes, HTML2IPF is the best solution, even though the code has to be adjusted via search and replace scripts and by hand afterwards. I think that there still is potential for further time-savings but to get there we first have to switch to XHTML/CSS and provide the authors with a style sheet that they can use in, e.g., NVU.

We would also like to point out that it may be a good idea to have a look at Wolfgang Draxler's improvements to Speedsoft Sibyl. At the moment, the most popular application that has been implemented with Sibyl is probably the INF viewer NewView.

---

March 30, 2005 - Michael Kramer didn't agree with the view our Editor in Chief outlined in the editorial of last month:

Your editorial provoked me to disagree. :-)

To me it seems that the problem [the low number of developers] does not only lie single-sided with the reluctance of people to contribute to software development. Looking at it more closely, I have to ask myself "What is actually missing?" There are browsers, office suites, compiler, and several other things that one needs. Of course, one gets the impression that OS/2 is treated stepmotherly but that is the result of IBM's disinterest and absence from the usual media. There is not much a software developer can do about that.

As someone who is quite capable of compiling, the question arises what "late entrants" can contribute at all. Many of the things that bother me I can hardly solve myself as I don't have the OS/2 expert knowledge of long years. (Actually, I am an Amiga and Unix user who came to OS/2 by a long way round.) Asking in USENET often does not get me there either. Either others are experiencing the same problem or I go on people's nerves because I'm the n-th person asking the same stuff. Googling often returns a large pile of crap of nothing at all, and the IBM web pages mostly seem to deal with migration of LAN environments but never with things that are important for developers. You can see that for many years, development for OS/2 has been unwanted by IBM. The toolkit is useful as a reference of functions but the thing we lack is an overview of how things are related and what to do with them. Something like a developer's bible like the Amiga Guru book or "Advanced UNIX programming." Or a (sorted) collection of source code one can use to see "how it's done." At Aminet, for instance, the sources are included most of the time, at Hobbes/Leo they aren't. Sure, there is edm/2 and many others, but it's all scattered, at random, and often half-baked like so many things (is that the reason they call is "half an OS?").

Where can I find something coherent and comprehensible about the following, for instance? (A random collection of questions that are open for me.)

• What is VIO and what can I use it for?
• How do I make GPI functions display the complete Unicode codepage on the screen? (Didn't find a single word about it, nowhere. Searching here and there brought up the undocumented codepage 1200 that brings the stuff from the symbol set to the screen at least.)
• How can you find out the current height of a string (in pixel, not font metrics)? (And, no, GpiQuery*() doesn't work.)
• What is the format of .ofm files?

There is a lack of "learning aids." And futhermore: of what use for the community is it if good-willed newbies write, e.g., drivers that are worthless because they don't have the necessary knowledge.

[Regarding C/C++ courses:] That is fine, of course, but knowing "plain" C/C++ is not enough for a complex system like OS/2, and you can learn it everywhere. What we lack is OS/2 specifics. Exactly the thing that is actually forbidden in times of platform-neutral development.

Christian Hennecke's answer:

We cannot change anything about IBM's attitude indeed. However, that is no reason to take the slow demise lying down and resign. In the end, everbody has to ask themselves if they want to continue using OS/2 and be able to continue doing so. Those who can answer that question with "Yes" will have to do something to keep "being able" at the same level or even improve it.

As far as the rest of your comment is concerned, I'm afraid that you did not get the message of my editorial. I completely agree with your statement that complete novices have a hard time learning the ropes of OS/2 development. Besides the things you mentioned, there are other obstacles that make the life of the beginner a hard one, e.g., the Toolkit 4.5x that can only be used after you have applied a number of modifications, outdated or hard to integrate compilers, etc.

But my point was not that hordes of end users should start with the development of new projects. Given that employees are usually blessed with little spare time, that would be a completely fantasy anyway. But it is possible for end users to contribute small bits and, on one hand, help advance a project slowly but surely, and even learn something on the way. On the other hand this will help relieve the more experienced developers who then can attend to things the beginners cannot implement due to lack of knowledge.

Experience has taught me that even as a greenhorn you can contribute something. In my editorial, I mentioned enhancements to the national language support of XWorkplace parts. Meanwhile I've managed to add a color configuration dialog to the XCenter Sentinel widget--almost without any knowledge of PM programming. A large step for me, a small step for XWorkplace, sure. But if enough others decided to do similar things, many projects could advance better. Continuing on the example of XWorkplace, it is relatively easy to add configuration dialogs for CONFIG.SYS driver statements.