“ScriptMeridian Archives: A Sob Story”

Last week, I promised the folks on the main ScriptMeridian mailing list that I'd have all of the archives (14,000 messages) set up in a browsable archive "soon".

Well, we're also planning the release of Conversant to the development community. One of the tasks that must be done first is a minor upgrade to the search engine. I decided that it was better to do the upgrade before indexing 14,000 messages, so I (along with a lot of help from Brian Andresen) spent the weekend and much of Monday on the AttSearchEngine.

Today I worked on the archives. I was given them in two different formats: a filemaker database, and a series of text files in Eudora's mailbox format. Both, I was told, were complete.

The database was more convenient (no parsing required), but it contains only subject, author, date, time, and body. Nothing to help with threading. So, I spent a few hours writing and running some scripts to import from the text files. This was slower, but a better approach because the message archives would browsable by thread.

When all of the messages were imported, what did I find? Only about 8800 messages, that's what. In other words, the text-based archive was incomplete.

That stinks! All of my work for the day was wasted.

Very few things in life irritate me as much as discovering that a challenging task which has been well done was a complete waste of time.

No more time to play with this, though. The archives just won't be threaded. C'est la vie.