Martijn Faassen wrote:
> Hi there,
> This is a summary of the previous discussions for those who weren't
> paying attention last week and don't want to read a huge thread coming
> back from vacations. I'm talking about you in particular, other steering
> group members. I'll spread it out over multiple short mails.
> So what happened?
> Hanno reduced the ZTK to a set of just zope.* packages. He removed the
> packages from the ZTK.
> I was happy with losing, but not the way the packages
> were just unceremoniously removed from the ZTK without discussion. I
> didn't like the way that was done, so I reverted Hanno's change and
> added the packages back to the ZTK trunk.
> I objected because of backwards compatibility concerns. Version
> information was being lost that was being maintained before. This
> information would be useful in both upgrading Zope 3 and Grok
> applications to the ZTK. (and I thought Zope 2 too at the time, but it
> is less of an issue there)
> People responded in various ways:
> * the ZTK isn't released yet and a discontinuity so there are no
> backwards compatibility concerns we need to care about.
> * people maintaining Zope 3 or Grok applications, or the developers of
> these frameworks, should figure this out for themselves.
> I responded that if the ZTK wants users it should help people cross the
> gap to it. That the ZTK has some responsibility to the transition from
> the past. Nobody seemed to feel that these arguments weighed very heavily.
> Since I'm writing this summary, I'm of course still right. :)
> Let me put that in other words: I believe have a legitimate concern that
> we could have solved pretty easily if we all weren't too busy trying to
> make points and being disgruntled.

+1 - I think concerns for backwards compatibility need to go beyond 
purely "academic" questions of what was released with what explicit or 
implicit promises, and consider what code is actually used in the wild. 
A lot of people are using* (in Plone for one). A lot of people 
need good KGS' to do *anything* and can't resolve their own dependency 
conflicts easily. A lot of people look to ZTK as the way forward for a 
sane, stable set of Zope packages to depend on. Those three groups of 
people overlap quite significantly. Maybe we never made them any strong 
promises, but that doesn't mean we shouldn't try to make their lives easier.


Author of `Professional Plone Development`, a book for developers who
want to work with Plone. See

Zope-Dev maillist  -
**  No cross posts or HTML encoding!  **
(Related lists - )

Reply via email to