On Thu, Mar 08, 2007, [EMAIL PROTECTED] wrote: > I occurs to me that any kind of KB needs real-life problems and solution, > must be user-driven, and have some form of unique problem identification. > > It seems to me that we already have a form of this in the squid-users > mailing list. But that one is very cluttered with non-problem messages and > with failed solution attempts. So it needs a bit of cleaning. > > What I am thinking of in a rough way is a KB which uses the squid-users > msg-IDs as the unique ID for a problem (after all the original problem has > been posted right?), and may in fact use the message itself from the > archive to describe the problem. Where a solution has been found, > references to the msg describin the solution could be used. > (By 'references' I mean either a web link by msg-ID or a PHP include to > show the archive content, descision to be made by the KB designer). > > Msg can continue to be posted to squid-users as normal, but the KB > sumbission 'authorization' being to link the posts to the KB. Rather than > ading an extra submission on part of the users, which is not that likely > to get used anyway. > > That would leverage an enourmous pile existing problem-solutions and > provide a framework for seamless and easy future additions. <dream>Might > even drop the clutter of repeated problem requests</dream>.
Thats not a bad idea - linking mailing list discussions back into KB articles - but the KB article does need to be a good summary of everything. Just needs a couple of warm bodies to watch the mailing list, jump on "problem situations" which crop up and could do with some documentation, and writing said KB article. I've made a start but we all know what my time availability is like.. adrian