Drudgery, horror, delight

It took a long harrowing day of debugging and rewriting elisp code capped by an hour of invigorating "OH CRAP DID I JUST DELETE 1100 UNREAD EMAIL MESSAGES?!", but I'm finally done - my office gnus setup with its last 3 years of email messages now works perfectly from home, connected to Wolfram's imap and smtp servers via vpn. Jamie S at the office tracked down my missing messages - it turns out that if you tell gnus to tell an imap server to split messages from INBOX to "imap-split", a file ~/imap-split is created with all the messages in it. 'G f' ahoy!

Along the way I finally read and understood The Fine Manual - it is possible to use an imap server as a simple message-serving pop server.

After the day's total immersion in elisp I switched back to mathematica to fix up some code and felt a bit of confusion, like trying to speak a foreign language I hadn't used for months.

Apologia pro code sua

It's rather cool to write dense convoluted Mathematica code to solve a problem, but the higher the density the harder it is to modify. Assumptions about the problem get superglued into the code; when the problem changes a bit, you have to go in and atomize the whole thing in order to change the code's basic assumptions. That's what I've been doing this morning with another guy's Mathematica code.

Or I could make the necessary changes by adding even more density and brittleness, guaranteeing that the next guy to work on the code will hunt me down and shoot me.

My coding style is distinctly uncool, like your Dad's station wagon, but all the assumptions are (usually) sitting right there out in the open in small easy pieces that can be changed in a couple of moments.

Wolfram Research roundup

• The Wolfram Demonstrations Project has been redesigned. Poke around a bit - there's some very cool stuff there.
• Wolfram's Mike Pilat has a short interview in the latest Dr Dobb's Developer Diaries.
• a new kind of "Hello, World" program in Mathematica at Walking Randomly.