"Joel on Software" on the Windows API

[jrtom]

"How Microsoft Lost the API War"

It's not clear to me that the future is as clear-cut in this regard as Joel makes it out to be. But it's an interesting article.

Comments

[red-frog.livejournal.com]

It's news to me that developers are no longer interested in the Windows API. Lots of interest in Web services, but when I was doing a regular column on them I could find no one able to say that they were using them, and there was little practical use for the column, I'm afraid. It was fun geeking, but that's all.

I think that Joel is reading MSDN too much and needs to talk to more people who still need to support legacy apps. And, for the record, I am looking very hard myself at the capabilities of the API and the enhancements going forward.

Incidentally, comparing upgrading to Longhorn from XP as being a less compelling upgrade than to Windows from DOS is, well, stupid. To paraphrase Pulp Fiction, it's not only not the same ballpark, it's not the same fucking sport. No, it's not the same, but there's a lot in there that will be very useful if they can make it work.

[jrtom.livejournal.com]

Fortunately (I always hated working with the Windows APIs) this is something that you're probably more clued into than I am these days.

As for your comment about the support of legacy apps, and web development: well, two things. One is that I think that he was suggesting that a lot of *new* development was web-based; if your column was a few years back, I would guess that things might well have changed in the time since. The other is that he seems to be implying that a lot of the legacy apps simply won't work under Longhorn, because the Windows API that they were written to use won't exist any more. Which seems like an impressively stupid thing for MS to do, but that seems to be what he was claiming.

I've heard that Longhorn is supposed to have some interesting new stuff going on. However, if they keep trimming major parts of it out (like the new file system) in order to make their schedule, then I doubt that it will have as much of an impact as MS would like it to.

[red-frog.livejournal.com]

I'm supposed to be up on it, anyway. :)

The thing is, some of it is Web-based, but there's also a ton you can do with the new interfaces. Avalon has some new stuff in it that ISVs are very interested in, but so far as I know is not alien enough to prevent today's apps from running. Yes, Web forms are cool, and, sadly, no one's doing much with them so far as I know. My query to readers of that column was about a year ago.

As for new things not working on Longhorn, too early to really tell, but that's not the plan as I understand it. For those apps that won't run (like the DOS apps that demand hardware access that NT OSes don't allow), we have VMs. An OS that will not support legacy apps is a dead OS, at least in the mainstream Windows world. A smaller market like Mac, selling largely to true believers, can maybe force the change better. Inertia in the Windows world is HUGE.

As for trimming, that worries me. A bunch of stuff was already supposed to be in R2 (interim release) and didn't make it. Then again, they could just implement all the stuff that was supposed to be cool about NTFS and never made easy to take advantage of, and they'd be better off.

Anyway... back to work. :)