Monday, December 31, 2007

Not Burke's Law, Brooks's Law...

(Boy, am I showing my age with that title...)

Fred Brooks will forever be known for having the misfortune of managing the development of the obscenity that is OS/360, and the genius to write one of the essential books about software development based on his experiences with that obscenity, The Mythical Man-Month.

I've referenced that book, and in particular Brooks's Law ("adding people to a late software project makes it later") in arguing against the "Open Letter" notion of bringing LL development to a screeching halt and putting everyone to work on bug fixes.

Other parts of that classic book also apply, though: "Plan to throw one away. You will, anyhow. The only question is whether you deliver the throwaway to your customers." (Apologies to all, and especially Professor Brooks, if I don't have that correct; I'm working from memory.) We as SL residents are currently using that throwaway. No offense to LL intended--I don't think that a system like SL can be designed all of a piece, perfect from the start. Not yet, at least. (To quote the title of another classic paper by Professor Brooks, there is "No Silver Bullet.") The question is, who's designing the non-throwaway: LL, or someone else? (Those of the "Open Letter" school are implicitly saying it shouldn't be LL, which will be the death of SL.)

That said, whoever is doing that designing had better also heed Brooks's warnings about "second system effect," in which designers, fresh from doing something the first time, all have their wish lists of things they couldn't do the first time... and throw them all in the second time around, to disastrous effect.

SL, for mass appeal, has to "just work." The average user simply will not put up with and doesn't want the collaborative environment that agile software development demands. So, who is willing to put up with it... and where can I sign up to do so? :)

P.S. If you look at Professor Brooks's curriculum vitae, you'll notice a paper on VR, and that virtual environments are among his research interests. I wonder what he thinks of SL?

No comments: