Knowledge Base Tree

[Victor STANESCU <bruno@Heineken.lmn.pub.ro>]

In order to corectly test the program i am working at would be nice to
have some clear enough ideea about the categories documents will be sorted
by. As I failed to find something like this on vodka, I have tried to
think about something. I would appreciate if someone could tell me whether
i am right about or at least correct me:

hardware
	video
	sound
	storage
		cd
		hard-disks
		tapes
		others(??)
	pnp(?? separate or merged with sound?)
	other categories..
system
	booting
	tuning
	networking
		services
		protocols
		client
	configuration
	installing (something for real dummies?)
applications
	programming
		ide
		compilers
		debuggers
		profiling
	office
	.. there are many things here, but well, even the thought of
		putting them all is scary.. may be you have found
		something smarter here

the current stage of the program is able to correct indicate the
apartenence to hardware/system/applications with a nice hit rate
(something like 7/10) for the documents in the Linux-HOWTOs directory.
i think it will be easy to parse html texts as all the problem seems to be
to ignore words in < > (html tags). but i want now to think about
something like a tree (after it puts the document in a category, enter in
a directory and something and read other configurations and place the
document in a subcategory of that category, and so on..	

Victor STANESCU-network administrator
The  Numerical  Methods  Laboratory,
Politehnica  University of Bucharest