[ZPT] spacify?

[Fred L. Drake, Jr.]

Marc Lindahl writes:
 > That's not a good criteria. There's a lot of legacy sites out there, 
 > and people don't change the code unless it's broken.  I have DTML sites 
 > running out there, and won't change them to ZPT unless I'm adding new 
 > stuff or doing a major overhaul....

It seems that having ZPT be compatible with DTML is not a criteria.
If you have a DTML site that works well and needs only minor
maintenance, why change it to ZPT at all?  DTML isn't going to
disappear in Zope 2 as far as I'm aware.

 > The design cycle is quick, but once things are deployed, the change 
 > cycle is slow... backwards compatibility is vital for zope to move 
 > forward.  If I start upgrading sites and *everything breaks*, I'll have 
 > to re-evaluate the efficacy if zope itself.

Agreed.  Your DTML sites should not break.  Your ZPT sites should not
break.  Switching a site from one to the other is a different beast,
as we're no longer talking about backward compatibility issues.

 > > There's been discussion of making it easier to use collections of
 > > helper functions by extending the path syntax; perhaps that would be a
 > > place to put functions of interest to DTML users.
 > 
 > or modules?

Things.  The idea has been to adapt something to some interface,
making a new set of callables available that operate on the portion of
the path that's been traversed so far, allowing additional
transformations to be performed.  String formatting operations are
just one set of interesting operations.

I'm sure you can dig back in the list archives for information about
the proposal.  I think there's a preliminary implementation on a
development branch somewhere, but I don't recall the branch name.


  -Fred

-- 
Fred L. Drake, Jr.  <fred at zope.com>
PythonLabs at Zope Corporation