I have a specific need for a debain dictionary -- I want to automatically define keywords in the Debian Weekly News so I can continue to use debian-specific terms therein without defining them; without confusing the larger readership of that newsletter.
I've seen http://www.debian.org/~elphick/ddp/manuals.html#dict and looked at the SGML source. I think what exists so far has a definite problem -- it's not well structured. I would like to be able to process the sgml source and get a list of all defined terms. I would like to then be able to link directly to a specific definition. I would also like to be able to have various metadata accociated with the terms. For example, how widespead is the term (do debian users use it, or just developers, or just developers in a subproject). What part of speech is the term, how is it pronounced, when did it enter the debian lexicon? For ease of automatic updating, it should be possible to add terms in any order and automatically generate html that is sorted alphabetically, with one page per letter as it is now. I think the debian dictionary needs a special-purpose DTD (which can perhaps just be some additional fields added on top of debiandoc), to allow this information to be included in it in a structured way. I'm thinking along the lines of an introduction/header similar to what is there now, maybe a similar footer, and in the middle have it source in a (possibly automatically generated) entity that contains the actual definitions. It also needs a well-publicised form interface to allow submitting information. The submitted items ould go live immediatly on a copy of the dictionary, and good ones can be added to the official dictionary. This should quickly get the dictionary full of definitions. So I am willing to put some time into this, though if I do any DTD work it will have to be in XML (I don't know all of SGML). Let me know. -- see shy jo
