Scott Hacker Responds

Elwood writes " Scott Hacker answers some of the questions and concerns that /.ers raised in response to his last column Of Tanks and Batmobiles."

Comments

On BeOS and Open Source

[Straker Skunk]

We all know what grief it tends to cause when you have a closed-source operating system with marketshare in the 90th percentile. I think many of us have seen this same potential, given the similar closed-source approach, in this newcomer to the OS show.

It's easy to imagine the "BeOS doomsday scenario"-- Be starts getting popular, then hits it really big, and then things start going wrong. Feature bloat comes in. The excellent API documentation starts slipping. New releases come out in model years. You know the path to the Dark Side.

If BeOS were open-source, we'd know what to do. Fork the code. Let the community keep them honest. But this, as things stand, is not to be an option for us.

Fool me once, shame on you. Fool me twice, shame on me. Anyone remember the old days, otherwise known as the '80's? I sure don't, but I've heard a lot about how Microsoft was the underdog back then. They were cheered on by many an end user against the behemothy likes of IBM, Lotus, and the other great Titans of old. Well, we know what happened to young Anakin. Is BeOS to become the next Dark Lord?

I don't intend to slight JLG, or any of the other intelligent and zany BeOS developers, whose efforts have proved nothing short of astounding. However, there is at least a basis for some, shall we say, healthy skepticism on the part of our more faithful open-source advocates.

We know Open Source isn't pixie dust, but it is a very good way of keeping a company from doing Microsoft-like things. BeOS has decided to keep their source secret. So far, they're doing this for pretty harmless reasons-- to run a tight ship, not tell the whole world exactly how the BeOS kicks a$$, generally reasonable things like that.

The problem is, at some point, it becomes very easy to do things with closed source that one really shouldn't do. Things like AARD, funky file formats, and talking paperclips come to mind.

I think the biggest fear around here is that, if Be does break critical mass and gets to the point where they can do that-- quite possible, in a post-Microsoft world-- there isn't going to be much we can do about it. Except kick ourselves for cheering Be on in their early days, call it BeO$, and have a Torvalds-wannabe write a new operating system that totally blows it out of the water. (After several years of careful development, of course).

Anyway, if Open Source isn't in place to keep them honest, I'm not sure what else could. Perhaps a cross-platform superset-of-POSIX API, such that the important apps can easily jump ship to Linux. (But then, given the differences between the systems, you either get an API that doesn't take full advantage of the BeOS, or that can't be acceptably implemented on Linux).

I'll definitely check out BeOS someday, but I do sometimes wish there were some tacitly acknowledged mechanism in place to keep Be in check, should such circumstances ever arise. RedHat has the GPL watching over its shoulder; TrollTech has the doomsday clause in the QPL. Be...?


My five cents. (keep the change)