Many years ago I wrote a blog entry wringing my hands about whether I'd ever get to write a Zope 3 application . The results are in, and it turns out that in the sense of being able to use Zope-3-the-application-server to regularly produce customer projects, I have not been able to, and I never will be able to. On the other hand, in the sense of being able to use Zope-3-the-libraries, I have, and I enjoy it immensely, mostly due to Martijn Faassen's and others' work on Five .
Despite Zope 3's maturity, Zope 2, fueled mostly by Plone, is still a far more popular platform for building web applications than is Zope 3 . Indeed, Plone's dominance as a platform has overtaken all Zopes .
"Zope 3" as an application server brand is dead. It's the same kind of dead as Mozilla Seamonkey as compared to Firefox. Or GNU as compared to Linux. While its useful and impressive libraries live on, the ability for its brand to inspire curiosity from significant numbers of new users is dead. That mantle has been handed to Plone.
The most convincing evidence of this comes from my consulting experience. It has become an abberance that I'm ever asked to do a "Zope project." Instead, I am almost always asked to do a "Plone project." For these, we often end up using plain-old-Zope2, but the lead always comes in through the Plone brand, never through the Zope brand. It has been this way since about 2005. Perhaps there's some bustling secret underground of trade in Zope 3 of which I'm unaware. I doubt it.
One of Martijn's other projects, Grok, which depends heavily on Zope 3 technologies, holds promise as a framework which will bring "pure" Zope 3 technologies to the masses. The king is dead, long live the king.