On Wed, 07 Mar 2007 16:35:25 -0600, Jim Wilson wrote:
Thanks Sheldon - those are excellent points. The conversion back-and-forth through HTML seems to be a sticky spot.
I wonder if there's any happy-medium? If pages could be flagged as "Wikitext" or "HTML", then using a WYSIWYG on the HTML pages and traditional edit on Wikitext articles could make sense.
Since the wiki text is more structured than the HTML, rendering it that way would be a lossy conversion, which is why it would be useful to have an intermediate XML that keeps the structure of the wiki text without its ambiguities.
Initially, I think this XML would be more useful as a transfer encoding, to export the full structure of the data to third-party software that doesn't need to emulate our Parser, and would take some time to mature before being used for storage.
Of course, we could consider wysiwyg to be be one of those programs, encoding what it understands into XML, and flagging what it doesn't, editing, then reversing it. So the medium could be a partial conversion, but done on the fly, and dealt with by the application.