Re: One Ring to Bind Them

From: Eric Kaun <>
Date: Tue, 15 Jun 2004 15:28:10 GMT
Message-ID: <eAEzc.25646$>

"Dawn M. Wolthuis" <> wrote in message news:cal5og$5e4$
> So, how should we fix the situation or is declarative vs procedural a
> of taste? --dawn

Declarative is better; there are enough different styles of declarative to satisfy many (not all) of those who find, say, Prolog distasteful. It's simply easier to lapse into procedural; you quickly find yourself in a quagmire, but that apparently is a lesson not easily learned (even by those who've been through one quagmire after another). It's that resistance to learning abstraction that's made me somewhat less tolerant of bad code than I used to be... the knowledge that in most cases, they'll just do the same type of thing again. And I'm even less intolerant of my own bad code... but until we use our procedural abilities to write "engines" that interpret declarations, we'll keep writing spaghetti. The trouble, of course, is that those focusing in procedural (and OO) generally don't see the value in bothering to deal with declarations at all.

What we're trying to accomplish is basic logic and computation; the restrictions of the languages we use, and the adherence to algorithmic thinking, keep us from advancing very far.

And, of course, the above is all just hand-waving and generalities, though generally true. I've just been debugging some horrific splicings of Java and InstallAnywhere (a rotten package with a GUI and no language at all), and am in a foul mood...

  • erk
Received on Tue Jun 15 2004 - 17:28:10 CEST

Original text of this message