Hey Sebastien,
Catalog tree FTW! thanks for the install!
check out the development catalog tree in action at the bottom of the page:
[objects.reprap.org]
That semantic mediawiki looks very cool, I think it would be worth trying
out on a test server if we end up bringing one online. Semantic wiki doesn't
look to be a replacement for Widgets - in fact they list it as a useful
extension to use with semantic wiki (
[semantic-mediawiki.org]). Some of the
functionality of Widgets I'm looking for could be alternatively obtained by
updating Media Wiki (so that we can have conditional statements in the
development template which i think are in a later version).
MW stuff: can you clarify? Why is mw bad for documenataion? What
alternatives are under consideration? Why are we porting to mw if it is not
where we want to be in the long term? I've read similar discussion on mw in
the forums but just assumed they were out of date.. If mw is not the end
goal then I may not be appropriately applying my efforts here in moving
everything over to it - I personally prefer MW but honestly as long as we
end up with documentation in one clean, accessible and concise location I'll
be happy.
Cheers,
Rob
Mediawiki is bad for RepRap for a few different reasons:
Here is a major example:
The developers instinctively run off with shiny things and tuck them away in their own private blogs and CMSs
[
dev.forums.reprap.org]
Repeat this several hundred times.
If we were a formal research collaboration under on roof, Adrian would sort out this sort of behavior with a few quite words in over tea. "We probably should document our research, you know?" We're not, so we have to build in blogging, and so on.
Also, blogging is more fun then writing anonymous docs.
Here's a really major example:
This is cutting edge research, which we can't support right now:
[
www.thingiverse.com]
I was extraordinarily frustrated when I saw this,
because this is the sort of thing that RepRap.org exists to support, but no one is using it, because it sucks.
And it's on a system which is probably going to be bought for USD5M by MS or Amazon two years from now, and then, *poof*, we've got zero influence on our CMS because the people who run it are too busy with their own projects to merge everything they're doing with RepRap.org, like, say a laser cut RepStrap or something. Or every extruder developed by the community in the last year.
It forks the community a bit. For inexplicable reasons, no one thought to set up the thingiverse software up as RepRap.org, with a bit of blogging tacked on, so ... here we are. It would have helped keep things as one community, rather than fork it a little bit.
Fundamentally, mediawiki is supposed to be a small kernel with lots of extensions. Community is bolted on using "talk", which is really fucked up on wikipedia - major nasty personal politics and catfighting. And all the real community-support for mediawiki is done on www.mwusers.com, meaning mediawiki can't eat its own dogfood.
Mediawiki is great for anonymous text articles and non-original synthesis of stuff trawled off the net.
RepRap.org is (should be) a space that is organized around parts, assemblies of parts, and discussions of those parts. Parts can be {extruder chunk, a RepStrap, or a sculpted figurine}, and the person who put it up needs to get the validation/critique/praise/discussion/etc. in order to want to do it. They'll want their personal page, and so on. And at the same time, we probably want articles+objects to auto-enter themselves into a reprap.org wiki as people blog them.
Vanilla mediawiki fundamentally fails this. We can bolt on extension after extension to do this, but eventually we're probably going to be utterly overextended and wishing we had started from scratch, like trying to turn a Ferrari into a cargo hovercraft, especially after we upgrade an extension and other things break, or we upgrade mediawiki to a new version, and our hand-coded glue and everything else breaks.
I think we do want to transition the twiki stuff to the mediawiki, but we need to keep our eyes on a postmediawiki solution which can track 10,000 users, 10,000 files, without us having to manually hand code into each page what the thing is. Because I think that would be the mediawiki solution.