Well, I finally found the time to get back to this glossary thing. With the help of the Highlight Search Queries plugin[?]by Tom Sommer, I was able to get some happy pattern matching for a glossary plugin.
This puppy is NOT FAST though, as it does regexp[?] pattern matching for each of the glossary words in the (optionally) entry bodies, extended bodies, and HTML areas on every page. I would reccomend only using it in the extended body, so the glossary plugin[?] won't have to do work on the summary blog views (assuming that your blog is setup thatway...)
The admin interface looks like the thumbnail at left (click for a bigger view). It allows the definition of a separator (the default is a colon) and a list of term, definition pairs. Each definition is ended by a newline (carriage return).
Additionally there are radio buttons to select which blog sections the markup should apply to, the Entry body, the Extended body, and HTML area elements.
The plugin uses a class, serendipity_glossaryMarkup to determin how the text and pointer icon will be rendered. A default class is created if one doesn't exist (thanks again, Tom), that is a brownish-red, with a help cursor pointer.
The plugin parses through the entry text and craps each incident of the terms with a span having a title set to the glossary definition. There is some effort to avoid changing html[?] code (for example, one of the terms I defined is span, which it ignores if in an html node <>. Also, the seach terms must be alphanumeric[?], with no spaces or symbols. If you don't like that, "use the source, Luke"
As stated,this is slow. My server is an antique PIII 400MHz, and running this plugin makes it crawl, so I am shutting it off on this site. If someone is curious to see how it runs in real life, drop me a line and I can re-enable it. In the meantime, the screen grab should give you an appreciation of the function.