Every time we get close to wrapping up another month's worth of DDJ, I start thinking of last February's issue with the multitasking juggler on the cover. What with last minute changes in the number of pages we have available, articles coming in long, articles coming in short, articles coming in late (or not at all), and editors in and out of the office at all hours, it sometimes amazes me that we get the magazine out at all. Add to that the desire for variety and balance among the articles (we don't want to have all C or all DOS articles) and we sometimes feel like we're juggling while walking on a tightrope.
As a case in point, this month we planned on running the follow-up to our February feature on the Rhealstone, our real-time benchmarking proposal. However, we underestimated the volume of response (and thanks to all of you who sent in comments) and we simply weren't able to evaluate and incorporate the recommendations in time for this issue. Our follow-up will run later this year. Likewise, in May we intended on running Martin Tracy's Zen Forth System, but we flat ran out of room. It too will run later this year.
As for operating systems, we were recently asked by a leading application developer what our cut -- or at least the cut of DDJ readers -- is on OS/2. Their reason for asking, they said, was that they have a lot of OS/2 products that (surprise, surprise) aren't selling too well; and a recent cross country trip hadn't shored up their spirits as major customers said that more DOS -- not OS/2 -- applications were what they wanted.
Our experience is that we're beginning to see a rise in the OS/2 noise level, although the major interest on the part of readers continues to be on DOS and, to a lesser degree, Unix. The difference in our experience and that of the company mentioned is that the noise we hear is coming from developers -- not end users who seem to be reacting with silence. We get calls pretty much every day from developers who are looking for OS/2 information. More significantly, we're getting more and more article proposals from programmers who've been working with OS/2. (One example of the latter is a nice little article by Nico Mak, who wrote April's popular SWAP article. After becoming frustrated with OS/2's cryptic error messages, Nico wrote a utility that provides a more detailed error report. Unfortunately, we again ran into one of those situations where we just didn't have room to run the utility this month, but you can expect to see it before long.)
Another thing we're hearing from developers who've been working with OS/2 is that once they've started programming under OS/2, they don't want to go back to DOS. That's the opinion of Mansour Safai, the chief architect of Logitech's MultiScope OS/2 debugger. (And if you get a chance to see this debugger, don't pass it up). I suspect that Nico Mak concurs even though he hasn't explicitly said so. What he has said is that he rarely, if ever, bothers to boot DOS anymore, OS/2 is his primary working environment.
What this leads up to is a growing demand for OS/2 products, but not the end-user applications people expected. Instead, the current market seems to be built around sophisticated development tools (like new debuggers).
The question developers must grapple with is twofold: should they spin their wheels working on OS/2 tools for a market that's still evolving, and how can they get those tools into programmer's hands. Vaughn Vernon's idea sounds good to me. Vaughn, who heads Aspen Scientific, thinks developers ought to give OS/2 tools away by bundling them with existing DOS packages. His theory is that developers can back into the OS/2 market with character-based tool kits, buying time to develop "real" OS/2 programs that have Presentation Manager support. Vaughn has been doing this for a while with his tools and I'm interested in keeping track of whether or not it works. What do you think?
Copyright © 1989, Dr. Dobb's Journal