Saturday, October 26, 2002

I Would Argue Like This. Language advocacy is usually (but not always) pretty boring to me. But this message from Richard Gabriel, who to me remains inscrutably smart, is just plain fun. No holds barred, man... [lemonodor] Oh, man. Follow the link. The next time I'm worried that I'm being too forthright and direct in indicating what I think about the software development tools we have to use to get hired, I'll just refer to this, and grin. Hmmm. Dick Gabriel is one of my intellectual heroes that I haven't yet met. I should take steps to rectify that. 8:41:50 AM    comment []   Lisp in Space. A couple months ago Erann Gat finished his personal history of lisp and other languages at JPL and Google (which may be a revised version of a previously noted usenet article). It's a somewhat tragic story that I find worth reading because It's a short insider's look at the sequence of robots that eventually culminated in the Mars Pathfinder rover. It sketches a not implausible case that but for some unfortunate politics, the Mars Pathfinder rover would have been running lisp instead of C. It mentions my old boss. It mentions the origins of OpenMCL. On the one hand, JPL has had some big successes since dumping Lisp. On the other, maybe they could have done much more if they hadn't been using languages that make software development almost uniquely difficult and slow. Even another ex-JPLer I know who thinks C++ is too far from the hardware says that JPL projects go sloooooowly. [lemonodor] Of course they do: JPL is government funded. C vs. Lisp has squat to do with it. Now, Dr. Gat is an acquaintance, and if he says that you can be 4-5 (at least) times more efficient in Lisp, then I believe him (and that also reflects my experience, but that might just have been the case for me). But Dr. Gat is also correct about the mental model of software development as a business: the idea absolutely is that developers are replaceable cogs, so all cogs must have the same morphology, which means that they must know the currently-trendy language (C, C++, Java...) My biggest gripe about the "personal" computing industry is that it went from invention to this level of stagnation in a scant 30 years. 8:35:44 AM    comment []   WebWork wins again. Struts meets Swing.  [Erik's Weblog] This article is totally ridiculous. When I read the introduction ("Most Struts applications use a browser as the client, but the framework is actually open enough to allow other client types.") I thought maybe Struts was as flexible as WebWork in allowing different types of clients to run actions. Reading on - this is definitely not the case. The application is done using the URLConnection class and sending data in plain text back and forth across the wire. Open enough? Bah! That's just being silly. [rebelutionary] I anticipate using WebWork in the future. It was first recommended to me by a colleague whose judgment about architectural matters I respect greatly, and since then I only continue to hear good things. The fact that it was adopted into the OpenSymphony effort only makes it that much more appealing. I can easily envision a trio of Resin, WebWork, and SiteMesh, perhaps with Hibernate for persistence, being a killer combo. 8:12:28 AM    comment []   IDG.  What happens when you mix $6.9 b, a flailing consulting company, and gung-ho security "experts" intent on taming how people use PCs.  An absolute mess.  If my math is even in the ballpark, this Intranet is costing the Navy (you and me actually -- no wonder we have a ~$150b deficit this year) ~$50,000 a seat.  [John Robb's Radio Weblog] Add to this the fact that you can't build a secure infrastructure around Windows, and the project was doomed from the outset. 8:05:58 AM    comment []

© Copyright 2002 Paul Snively. Last update: 11/5/02; 5:55:05 AM. This theme is based on the SoundWaves (blue) Manila theme.