[Zope-Coders] Proposed 2.7 features

[Jeffrey P Shell]

On Monday, January 20, 2003, at 09:39  AM, Brian Lloyd wrote:

>> I made three proposals for Zope 2.7 and I would like to hear if they
>> can be merged into the HEAD. Since I will be very busy until 
>> mid-February
>> I would like to finish the work as soon as possible (early before
>> the alpha release).
>>
>> Andreas
>
> (Sorry this is so late - I've been buried)
>
> I think the index names and DateTime proposals are good to go
> (note I made a few comments on the DateTime proposal regarding
> unit test coverage and docs).
>
> The ReST proposal makes me a little less comfortable, since it
> provides "more than one way to do the same thing", which is
> often a signal that something should remain an add-on. (FWIW,
> Jim's opinion is that believes it was a mistake to tightly
> integrate old-style STX with DTML documents, etc. in the first
> place). It seems a little thing, but IMHO it is an accumulation
> of these types of little things that contribute to some people's
> opinions that Zope is inscrutable.
>
>
> That said, I'm willing to defer to the will-o-the-community on
> this. I do think that there needs to be good docs and tests
> ready before it is merged. Docs need to include updates to the
> relevant DTML parts of the Zope book, updated help for DTML
> objects and some sort of ReST reference (though I imagine this
> could be a link to the official ReST docs).

As much as I really really really want ReST integrated into Zope (I've 
had more success using it for documentation efforts than I've ever been 
able to achieve with STX/STX-NG), the ``docutils`` package is still 
young and going through lots of changes.  The ReST format itself is 
pretty solid, but the internals of the parser are still going through 
changes with some new node types being added.  The last official 
release was 0.2, which was quite some time ago.  A lot of changes have 
happened since then.

So as it stands, adding ReST support could be a configuration 
management nightmare as long as "nightly builds" are the preferred way 
of getting the docutils code.  I'm willing to get involved with the 
docutils development lists to get a better idea on their plans for 
future releases and release management.  But due to this fact alone, 
I'm leaning towards the idea that ReST should be an add-on.

Perhaps a good alternative proposal would be for a generic "Formatted 
Text Handling" service that DTML, CMF, etc, could use, so that someone 
could write a ReST Formatter and existing objects could make use of it. 
  (The same goes for "WebText" as used in CMF Collector).