*blows dust off list*
90% documentation of wikitext is not hard. The last 10% has killed every attempt.
[percentages not exact]
So. I suspect what will happen to get WYSIWYG editing is that FCKeditor or similar will get a fair way into that 10%, then whatever remains will be deprecated and bot-converted away from.
Then we translate FCKeditor into EBNF/ANTLR/WTFBOL and the world is a utopia of third-party parsers galore and everything is rainbows and unicorns!
The question that springs to my mind is: What is that last 10%? Has anyone documented the pitfalls that killed the intrepid explorers of the past?
- d.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi
*blows dust off list*
*cough*
90% documentation of wikitext is not hard. The last 10% has killed every attempt.
[percentages not exact]
So. I suspect what will happen to get WYSIWYG editing is that FCKeditor or similar will get a fair way into that 10%, then whatever remains will be deprecated and bot-converted away from.
The biggest problem with that idea is that you can't draw a clear line between those 90% and 10%. All wikitext is valid by defintion. How do you know if you parsed something correctly or not?
It isn't a single construct that's problematic. It's corner cases with very common constructs. You can't deprecate {{...}} and {{{...}}} syntax (which at least in my experience are the most problematic).
Also these cases can't be bot-converted away for the same reason third-party parsers and WYSIWYG editors can't candle them.
Then we translate FCKeditor into EBNF/ANTLR/WTFBOL and the world is a utopia of third-party parsers galore and everything is rainbows and unicorns!
I don't know about FCKeditor, but I did bring Wikiprep parser very close to full compatibility with MediaWiki and I can say that it would be no easier translate to a formal grammar than the original.
The question that springs to my mind is: What is that last 10%? Has anyone documented the pitfalls that killed the intrepid explorers of the past?
Check the archives of this list. It appears that every 6 months someone will appear and try to find a solution (I learned my lesson a year ago and I've been pessimistic about this problem ever since).
Regards Tomaž
- -- Tomaž Šolc, Research & Development Zemanta Ltd, London, Ljubljana www.zemanta.com mail: tomaz@zemanta.com blog: http://www.tablix.org/~avian/blog
wikitext-l@lists.wikimedia.org