wiki-tr

I'm trying to think of some changes to wiki-tr which will address some
of its weaknesses.  I'm interested in feedback.

   1.  The current mode, where it produces a document on-demand and
       rather slowly, will be called "Preview" mode.  The Preview will
       have some information at the top to explain that it's a preview,
       maybe some in-line error reporting, and which will include a
       "Publish" web-form.

   2.  If you like the preview, and are logged on to the website, you'll
       be able to "Publish" the draft, which will make it appear as
       a normal (fast, stable-forever) document on w3.org.  

   3.  I'm trying to figure out if the publication name should be
       automatically generated based on the date:
          http://www.w3.org/2005/rules/wg/drafts/core-20070322b
       or based on a simple version number:
          http://www.w3.org/2005/rules/wg/drafts/core-v122
       or based on a version series (like software releases)
          http://www.w3.org/2005/rules/wg/drafts/core-v0.3.1
       or just left up to the user (who might follow one of the above
       approaches, but would probably have a hard time doing it
       accurately.

   4.  I'm trying to figure out if there are diffs I can automatically
       publish that would be useful.   For instance, when I was working
       on the RIF charter, lI inked each draft to a History page, which
       had links to each "significant version" and the changes between
       them:

          http://www.w3.org/2005/07/rules/charter2_history.html

       (Notice the red and green text on the "changes" documents)

       I could do that automatically here, but often what users wants is
       not a diff from each change to the next, but rather a diff from
       the last version they read!  The Wiki has arbitrary
       change-comparison code, but I don't think I'm up for
       re-implementing that.  Maybe I could just make a web-form where
       you enter the two URLs you want to diff.  One possible middle
       ground, using release versioning, is to say version 1.3.4 would
       include diffs from 1.0, 1.3, and 1.3.3.  That assumes people are
       likely to read the more major releases.  (I'd say 1.0 = first
       public working draft, 2.0 = second public working draft, etc, and
       the .n releases would be "editor's drafts" which the WG is
       supposed to review.  Imposing that structure on drafts might help
       keep things clear.)

Thoughts?

       -- Sandro

Received on Thursday, 22 March 2007 15:37:48 UTC