Each document should be indexed with, and searchable by, the following metadata if at all possible. See Definitions#Terms for what these words mean.
TODO check out the Dublin Core list and grab what makes sense.
Possible terms
- title
- source ID (GUID? SHA256 of a canonical name?)
- world
- author(s)
- Date created; date last updated (to the extent we can tell)
- Format (text, video, magazine, book, ...)
- For folks just getting started, should we also index links to, or contents of, free asset libraries? I think coders reading ps's 14days would benefit from some gfx and msx to start with.
- AV flag indicating the doc discusses graphics or sound. E.g., generic "intro to C" docs would not have the AV flag set, and the default search would exclude such documents.
- Handles (doi, ISBN, ISSN, URL, other name or location info)
- Subject area (code/gfx/msx)?
- Expertise level of the authors?
- Expertise level of the target audience (is this document intended for newbies or experts?)
- Detail level of the document / how close to implementation (e.g,. for gfx, is it just the math, or does it also have code?)
- Type (introduction, topic-specific tutorial, walkthrough, course, reference, tool docs, annotated source?)
- ps's 14 days and gargaj's codeacademy fit somewhere
- Code license?
- Maybe just a commercial vs. noncommercial flag so you can filter out results that use expensive tools
- Applicable tools (VS, gimp, openmpt, ...)
