On 8 February 2012 15:06, Oren Bochman orenbochman@gmail.com wrote:
I think that an xml subset is the ideal should be the underlying format. It's the best known technology, has mature development tools. It could be parsed to and written to most efficiently by browser, and even the editor could be simplified by using it.
This is likely. But the magical key to the issue is: it doesn't actually matter.
The important thing about the eventual intermediate format is that it will be *properly defined* and can be manipulated. It will be things that are processed in a particular way.
This format could be stored in whatever format is convenient - XML, bytecode, cat GIFs, a version of the existing wikitext format - as a properly defined syntax, we can arbitrarily transform any version into any other version as needed.
Hopefully this is possible ...
- d.