Michel Pelletier wrote: [snip]
This is such a good point I wanted to share it with the Zope community as well. The language issue is minor to the need for a definate object model with clear interfaces. It shouldn't matter if the core of Zope is written in Python or FORTH, from your perspective and the perspective of your language the model and interfaces should be well definded.
While I completely agree that identifying clear interfaces and object models is extremely important, it does matter what the core of Zope is written in, and not just for DC. I *like* it that Zope's sources are written in Python, and not in, say, C++, or Perl, or Forth. That's because Python's far more readable to me than those languages. This means that reading the source can aid me in understanding, and I can potentially go in and change stuff. Now of course we may say that, with clear documented interfaces and a clear documented object model, I'll need to go into the source far less, at least to gain more understanding. That would be nice of us to say, but then we'd have to look at the current state of the documentation as well as the current definition of the interfaces and object model; they're rather lacking. So, while I agree we should strive for these things, we should also not forget that implementation languages *do* matter in the real world, and due to the behavior of the real world and real people, this won't be changing anytime soon. Don't-forget-open-source-ly yours, Martijn