Most new products are upgrades of old ones. I don't mean in the senseof upgrade sales (although that too), but in terms of the code they'rebuilt from; they are revisions of existing codebases, not freshlywritten from scratch.

This has extremely rational economic and developmentreasons, as articulated by Joel Spolsky amongothers; there is a lot of money and value in existing codebases. Butignoring all the arguments about cost and captured knowledge and soon, there's another big reason

Recently, we had an interesting fire drill that actually winds up beingthe first decent argument for periodically rebooting good servers thatI've seen.

We have a number of very important servers, the kind of importantservers that are active all the time and that have their downtimescarefully scheduled well in advance. Recently, we (and by this I mean 'aco-worker') had to patch the OS on a couple of them in the pack. Thiswent fine.

As part of patching the OS,

The SMTP listener crashed and was restarted around Wednesday at 2am,so some of the statistics are short this week. That said, this weekwe:

• got 12,378 messages from 234 different IP addresses.
• handled 17,251 sessions from 822 different IP addresses.
• received 87,872 connections from at least 29,223 different IPaddresses since Wednesday at 2am.
• hit a highwater of 6 connections being checked at once, since

One of the under-appreciated areas of writing systems documentation istesting it. And I don't mean proofreading it (although that's necessarytoo); I mean making sure that it is things like comprehensible, clear,correct, and complete.

Testing is vital, because documentation that is incomplete or incorrectis often worse than no documentation at all. Documentation createsconfidence, which means that bad documentation lets people go wrongwith confidence, turning a small mess into a much bigger one. ('Butthe