Wednesday, March 22, 2000
Searching for the unsearchable
I'm attempting to find a list of “noise words,” word to avoid indexing for a search. I found plenty of references to noise words, but no actual list. Half an hour I spent, and then it hits me—would a web page that lists typical noise words actually be indexed?Hmmmmmmmmmmmmmmm indeed …
If you think sendmail.cf is bad …
Mark and I ended up helping John (the dot com paper millionaire and keyboardist for a Grateful Dead coverband) set up a new feature on his site. It involved deep hacking of both sendmail.cf and the Sendmail source code itself.Some things we found out:
- Sendmail ignores the current environment setting, using its own environment when it runs.
- It's easy to see why sendmail has so many buffer overwrites—the code is attrocious.
- Sure, sendmail now allows you to name rule sets instead of number them (in sendmail.cf), but does it show the name when running in address test mode? Nooooooo …
- Network analyzers are fun. Very impressive the way they can determine cable length.
- Eric Allman has a lot of explaining to do.
We might have actually finished had John not crashed on us.
