Which tags do we allow?
This is what everybody does right now. Just every combination of letters will be valid as a single tag.
- Different people might understand different things under the same terms, but maybe this is irrelevant if you search for more than a single word?
- SPAM ..
Allowing only words that are title of a Wikipedia-Site, or at least use wikipedia as follows to improve the quality of tags automatically:
This could solve some problems:
- Words that are acronyms with several meanings could be solved via the information that is contained within wikipedia already. (e.g.: http://en.wikipedia.org/wiki/DRM) The tagging-interface then could automatically provide a list to the user and allow to choose one of those meanings (by replacing the tag with the correct wikipedia article name).
- Each tag has a described meaning, thus people will more likely understand the same
- A search through tags would provide the searcher the same lists for which meaning (of such a word) to search.
- Often there are several words that mean the same thing. This can also be automatically solved with the help of Wikipedia: Redirects. Redirects are Wikipedia-Sites that contain nothing but a reference to the article with the correct lemma for its meaning. Such tags could be automatically replaced with its correct lemma. (e.g.: http://en.wikipedia.org/w/index.php?title=Asymmetric_Algorithms&redirect=no)
- some words might be missing completely ..
Lots of other good ideas about a Software Thesaurus can be found here: http://www.groklaw.net/article.php?story=2006052522234947
- foldoc (http://foldoc.org)
- NAG (http://www.nag.co.uk/numeric/fl/manual/html/FLlibrarymanual.asp)
- UNESCO Thesaurus (http://www2.ulcc.ac.uk/unesco/)
- DADS (http://www.nist.gov/dads/)
- IEEE Standard Glossary of Software Engineering Terminology (rev 2002) (http://standards.ieee.org/reading/ieee/std/se/610.12-1990.pdf) -> this one also sound promising but I have no account there so I couldn't take a look at it :-(
- Linux Dictionary (http://www.tldp.org/LDP/Linux-Dictionary/html/index.html)