[Zope3-dev] Re: Reflections on the Zope 2 to Zope 3 transition

[Joachim Werner]

Hi! 
 
> 2.7 definitely took too long which is not so much ZC's fault. 
 
I'm not sure what's going wrong with my projects, but apart from the 
"SessionGeddon" a couple of years ago (Session API changing between beta and 
final without anybody telling us) switching to a new Zope version hasn't 
really caused me too much headache. 
 
 
> > Before I get into the main topic of my response, I'll note that I'm 
> > good bit more optimistic about backward compatability than you are. 
  
> Good for you. I'll rather stay pessimistic and be positively surprised 
> in the end, maybe. 
 
I am quite positive that it will be easy to port Zope2 content objects to 
Zope3 systems without too much effort. I am also quite sure that porting the 
actual code will be a bit harder. But OTOH either you don't like the old code 
and want to rewrite it anyway or you can just keep it and be happy using Zope 
2 for another couple of years ... (so this is no real issue for me either). 
 
> > 2. I decided not to think much about the transition until Zope 3 has 
> > gotten farther along. 
  
> I guess this was the main thing I do not feel comfortable with. I'm not 
> saying you decided wrong. I'm just wondering whether the current  
> situation could have been avoided. 
 
Hard to say. I've already said before what I think we'd have needed: A 
concrete example for what Zope 3 should be the toolkit for. Most really good 
frameworks are good because there is a real-life application that uses the 
framework. Take Qt. Without KDE there would not be half as many users of Qt, 
and getting the rough parts fixed would have taken much longer than with KDE 
being the main driver. OTOH Qt is a perfect example of a good architecture to 
start with. So maybe the "KDE" for Zope 3 is still ahead ... 
 
Cheers 
 
 
Joachim 
 
 
_________________ 
 
Wittelsbacherstr. 23b 
90475 Nürnberg 
 
joachim.werner at iuveno-net.de 
www.iuveno-net.de 
 
Tel.: +49 (0) 911/ 9 88 39 84