2007-08-29: w00t! OpenOffice translator in the works!
Man this is awesome. Strykar was trying to write an article, and he was having problems with the Cowpoke engine accepting it - we later figured out it's a character encoding thing... I guess that things like long dashes and stuff are acceptable in their non-entity forms in UTF-8 encoded XML. For some reason, however, the parser cowpoke's using (at the present time, the one built into PHP) is assuming us-ascii encoding, regardless of what's in the XML header. Sucks.
Anyway, I started goofing around with the XML filter packages for OpenOffice, and found that with just a few hours work hacking up the .xslt files from the DocBook filter, I was able to chop together a filter that'll output XML that Cowpoke speaks. Almost.
Unfortunately, I haven't figured out how to get it to output any other kind of encoding than UTF-8. This is a problem, because any dashes (such as the one in the first paragraph) and/or weird apostrophes make it choke. I have to go back through by hand and replace them, and change the encoding as well (or just remove it). Hopefully I'll figure it out though, and release the filter package so that Strykar may use it. :D
Once I sort out the character encoding, there's also the problem of how I'm going to handle images, and things of that nature. I'm trying to keep the customizations to OO.o to a minimum, but it might turn out I have to release a template with the filter package as well. We shall see.
The final hurdle will be writing a translator that goes the other way - allowing you to download a file from the server, edit it, and re-upload it. I mightn't ever bother with this - I might just wind up implementing DocBook as the arse-end of Cowpoke anyway, I still haven't figured out which will work best.
The carnival's all set up, they open tomorrow at 5pm. Looking forward to it. :D