There really isn't much to talk about this point as the project is only slightly more involved than proof of concept and what is here well, it's probably not terribly useful to anyone who's not purely a native english speaker.
Like most OSS, I had a task I wanted to accomplish and found the existing solutions wanting. This isn't to say that I had a particularly strong clue about how to do any of it, but it's a start.
What am I trying to do? I personally find that having more than 20 or so face installed at any one point make it inceasingly difficult to find the face I want to use in a given circumstance. In the past, on Windows based systems, I had used applications that me assemble collections of faces for particular purposes and activate them or deactivate them as needed. This, I found, was the best way to solve my particular problem.
Of course, it wasn't perfect, nothing is. The applications I used were deoendant upon the fonts themselves for any categorization aside for assembling them into collections. This was non-optimal for a number of reasons, from a lack of font included declarative information ( script/serif/etc ) to the arbitrary categorization.
At this point, there are proof of concept scripts and binaries that build via a VERY crude sCons script. In theory it also builds and installs a linkable library but this has yet to be tested and is FULLY based upon my own system's paths. But it does work with hand built configuiration files.
Sadly, an extensive list. I'm a Pre-C89 self-taught coder. I have an astounding variety of bad habits from those days that seem to drive modern open source hackers bugnutz. If you grew up on a mix of K&R, Mixx PowerC, and TurboC 2.0 you'd have some issues too. :) Questionably named variables, silly comments, and *gasp* no objects to be found! If you'd like to contribute, please consider that this is as much a learning exercise as a project - I appreciate the help but I would find advice and explanations of why things are improvements a lot more useful that "The old way sucked, this is better!" and thus be far more likley to commit the change. I won't commit changes I don't understand, if you can't or won't explain then don't waste our mutual time.
The current setup doesn't even pretend to be international, it's barely ASCII7. It only handles single file typefaces.
About my terminology - Like everything else here, it's self taught. This makes it a mix of bad habits, incorrect, and outdated terms. So until I get around to learning the correct modern vernacular of computer typography, I'm using these terms the following ways.