[Zope-dev] The bleak Future of Zope?!

Terry Hancock hancock at anansispaceworks.com
Wed Apr 21 10:39:19 EDT 2004


On Wednesday 21 April 2004 05:52 am, Eckart Hertzler wrote:
> I don't agree.
> I am new to zope. So I tried zope2 first, because plone had a lot of appeal.
> I got discouraged very quickly, because zope2 is so very grown over a time
> it's hard to join later.
> 
> Zope3 seemed quite well documented and I had no problems going on on my own.
> ( There is a tutorial, a cookbook, and an online apidoc )
> 
> I can say nothing however to migrating apps from zope2 to zope3.

I'm really looking forward to Zope 3, and I'm thinking about migrating to
it this Summer.

I've been developing an application, which has taken about two years, largely
because developing in the Zope 2 "Framework" model is like beating your head
against the wall constantly.

That's probably because I'm writing a fundamentally complex web application
which I need to have a lot of large-scale control over.  I'm not writing in
an environment where a "slightly-customized" ZMI or even a "collection of new
Zope objects" will quite do the job. I'm writing a system which gives end-users
(NOT CS experts) a lot of control over their environment.  And there are 
fundamental user-interface changes involved.

I also have to do this in my "copious free time", as I'm not commercially
employed to do this work (maybe someday, but not now). So in those two
years, I've probably had the equivalent of 2 months of full-time work.  For
somebody dealing with that, the constant pressure to adapt to a changing
platform and the myriad interfaces that break when you do, and the
unwillingness to document these problems "because that's too old" get
really frustrating.  The lack of formally defined interfaces makes
it very hard to deal with this situation -- it's not easy to mix-and-match
the new parts you need with the old parts you haven't been able to
upgrade yet.

In short -- Zope 2 is TOO LABOR INTENSIVE.  Mostly because it's TOO COMPLEX
and TOO MONOLITHIC.  During the development phase of my project, I've had to
upgrade Zope THREE times, and EACH one REQUIRED A MAJOR RE-WRITE on my part.
That makes it very difficult to concentrate on forward momentum.  I've missed
my own deadlines, and had to admit that I simply can't deliver the product
on anything like the schedule I originally was trying for.  And this "3 steps
forward, 2 steps back" problem of dealing with a changing, poorly documented,
and often buggy platform is part of the reason.

The promise of Zope 3 is that it is following Python's TOOLBOX model, and making
it easier to separate out the parts you need into separate interfaceable
components.  This will make life vastly easier for large-scale projects which
don't follow the typical "quick and dirty" Zope site model.

Or so I hope. ;-)

I don't understand everything else they're doing with it, and I've had frustrations
with Zope 3, but in the long run (which I care about -- I expect my application,
or a later version of it, to be in use in 15-20 years, so I'm not just concerned
with "first to market"), I think it will be easier to keep up with.

I understand that my situation is probably unusual, but I do want to speak out
to say that there is interest in Zope 3, and I personally expect to be using
it before 2005.

Cheers,
Terry
--
Terry Hancock ( hancock at anansispaceworks.com )
Anansi Spaceworks  http://www.anansispaceworks.com




More information about the Zope-Dev mailing list