TUMBOODO: What was I thinking? [0730.1998]

Michal's first rule of database design: have a healthy supply of sample data. Simple, huh? Any idea why I didn't do it?

The data structure I'm currently using isn't nearly powerful enough to handle the complex relationships between different programming languaes.

I've been acting as if overlap between languages was all or nothing for any particular node in the code tree: VBscript and Javascript have a lot of overlap with each other, but not much with, say, SQL.

For one thing, there's all kinds of declaritive languages I want to cover: SQL, Prolog, HTML, XML, etc.

Also, 10 different topics in one language might all map to the same topic in another. For example, C's pointer operations should probably map over to a single Perl topic explaining the perlian idea of references, and have links from there to the various reference-related routines.

I could have handled this months ago if I'd simply collected the data for two or three languages.