On Thu, Mar 20, 2008 at 12:44 PM, Cormac Lawler cormaggio@gmail.com wrote:
- Would this issue be solved quicker by having a few funded developers
working on it? (If so, how many wo/man hours would be needed - what size of a funding bid would be needed?)
It could be partially solved if people were working on it. And, well, it is. Wikia has an extension, for instance, and there's FCKeditor. I at least have generally assumed that it can't be completely solved (i.e., having feature parity with wikitext editing without common or unfixable bugs) without a fairly dramatic rethink of a large part of MediaWiki's infrastructure. Whether such a rethink would be accepted into core to begin with is a fairly important question that seems unanswered.
As for man-hours (what, you've never read The Mythical Man-Month?), it's probably closer to three months of work than to three days or three years to write a complete solution, assuming it would be accepted. More precise than that I'm not going to attempt. Of course this assumes it's done by somehow who really knows what they're doing. And it assumes a "proper" solution is necessary.
- What does the fckeditor in existence *not* do yet? <
http://mediawiki.fckeditor.net%3E (If the answer is: "too much - it only does a tiny handful of things compared to what we need" - well then, that answers my question. :-))
I haven't looked at the tool much. It's almost certain that there will be some cases where it will fail horribly. However, glancing it over, it looks very slick and well-done. It loads up [[George W. Bush]] perfectly, to pick a representative heavily-marked-up article, except that it doesn't seem to display the images right. It would be very interesting to try enabling it experimentally at Wikimedia as an option to see if we get any complaints about non-idempotent saves or anything else problematic.
I'm still sure there are some fuzzy spots where it will break, but as a partial solution it seems better than I had expected. It would be interesting to consider the possibility of eventually including it in MediaWiki core, or shipping it as a preinstalled extension. The first step for that is going to have to be testing on Wikimedia sites, I think, which needs one of the roots to come on board, like Brion. What are your thoughts on trying it out (as a disabled-by-default option), Brion?