Ian said:
Because everyone has made their own, and no one can agree on what a
dominant web framework might be. Zope *was* that framework, with
dominant mindshare, but it didn't keep other people from making other
frameworks.
There's already a ton of frameworks, but I think I saw two brand new
ones on PyPI this last week. Crazy.
Maybe the current web frameworks are just not good enough? Also, there
are many different perspectives on how to do web-applications or
generate pages.
The best i hope for is factoring out common functionality into small
modules which eventually make it into the standard library. It's
not easy to get there, though. Analyzing the existing frameworks
down to the module level would be a good start, i guess.
Maybe the cost of sharing and reusing code is too high. If you e.g.
want to use a module from another project but not the whole thing
it's inconvenient and fragile to do this. Actually the Zope3-developers
tried to refactor common functionality into independent packages
(ZODB, the TAL-interpreter, security etc.) but i am not exactly sure
how well this will work out. It's still a huge thing with many pieces.
I like to think that the problem lies with the *framework* character
when i really just want a *library/module* that i can use. Frameworks
tend to impose a specific view of the world you have to adapt to. Thus
many people decide to roll their own ... often resulting in the
same problems for others. It's a vicious circle.
just my 2ec,
holger