thoughts on using the mediawiki "visual editor"

44 views
Skip to first unread message

Jim Garrison

unread,
Aug 11, 2012, 10:59:30 PM8/11/12
to wiki...@googlegroups.com
We've been talking for a bit about ways to make the Wikiotics/ductus
textwiki easier to edit. Our ultimate goal is to move to a visual
editing system (not requiring the knowledge of a special syntax), but
without requiring large amounts of development work on our part.

In particular, one thing we've discussed is trying to piggyback on
mediawiki's "visual editor" which is in development. Wikimedia
maintains a mailing list ("wikitext-l") for discussing issues with
parsing and visual editing, with archives at

http://lists.wikimedia.org/pipermail/wikitext-l/

I would characterize the following thread as a lukewarm reaction to the
idea of making the visual editor work for things other than mediawiki.
(This is slightly different than what Ian heard on the ground at
wikimania, but I am inclined to trust it.)

http://lists.wikimedia.org/pipermail/wikitext-l/2012-May/000712.html

Also of interest from this thread:

> VisualEditor is the front-end, it does not parse Wikitext, it reads in and
> emits out annotated HTML

The annotated HTML is (from what I understand) translated to/from
wikitext by Parsoid, the parser that the mediawiki team has been working on.

In particular, we need to find out what "annotated HTML" means
specifically in the above sentence. Is it well-defined yet, or is it
something that's in a state of flux? Is it something we can hook onto
(and use as a storage format for our own wikitext), or is it going to be
something tightly coupled to the intricacies and needs and mediawiki itself?

I am reading through the following thread from six months ago, which I
believe is the most informative public discussion of these issues so
far. As of then, the proposal was to use some sort of HTML5 microdata
to store annotations, but it does not seem that anything was finalized
during that discussion.


http://lists.wikimedia.org/pipermail/wikitext-l/2012-February/000517.html

Laurent Savaëte

unread,
Aug 13, 2012, 11:55:55 AM8/13/12
to wiki...@googlegroups.com
There is some interesting material in the list indeed.

The following two pages give some extra insight

http://www.mediawiki.org/wiki/Future/HTML5_DOM_with_microdata
http://www.mediawiki.org/wiki/Parsoid/RDFa_vocabulary

and list a few equivalents between wikitext and "annotated HTML" (at
least, from what I understand). The html looks quite heavy, but if it
sticks to the microdata standards, then it renders directly in browser,
which I find quite nice! (because speaking standard HTML/DOM means
easier development for tools etc...)

There is also a thread about dumping the original assumption that the
content of a page is wikitext
http://lists.wikimedia.org/pipermail/wikitext-l/2012-March/000661.html
and adding some content type info, so they're catching up ;)

Jim Garrison

unread,
Aug 14, 2012, 12:26:15 AM8/14/12
to wiki...@googlegroups.com
On 08/13/12 08:55, Laurent Savaëte wrote:
> There is some interesting material in the list indeed.
>
> The following two pages give some extra insight
>
> http://www.mediawiki.org/wiki/Future/HTML5_DOM_with_microdata
> http://www.mediawiki.org/wiki/Parsoid/RDFa_vocabulary

Thanks for finding these links. Things look clean enough that we could
conceivably use that proposal for our wiki markup language directly.

> and list a few equivalents between wikitext and "annotated HTML" (at
> least, from what I understand). The html looks quite heavy, but if it
> sticks to the microdata standards, then it renders directly in browser,
> which I find quite nice! (because speaking standard HTML/DOM means
> easier development for tools etc...)
>
> There is also a thread about dumping the original assumption that the
> content of a page is wikitext
> http://lists.wikimedia.org/pipermail/wikitext-l/2012-March/000661.html
> and adding some content type info, so they're catching up ;)

I saw that! :)

- Jim
Reply all
Reply to author
Forward
0 new messages