Am 07.03.2006 um 14:42 schrieb Martijn Faassen:
We need a way to enable the Zope 2 TTW person to be productive
without that person creating huge maintenance costs later on. Code
created by ZMI developers should look like something that could
also be on the filesystem, and can be checked into SVN, and
refactored and developed further, i.e. be taken into real
maintenance without major hurdles.
I would identify these functionalities, to give people something they
had in part through TTW.
1. Nearly zero setup:
Just edit a function with a context given as parameter. That's
actually at the base of an adapter.
2. Easy determination of context or locality:
The nice thing with scripts is, that they can be placed somewhere in
the content tree. But the meaning for the developer is, that they do
not know much about interfaces, he simply sees the spot it should be
used. Helping the developer to say "here do this" would be a big
benefit.
3. Scripts only get value, if they can interact with a more powerful
object api. This actually marks them as parts of something
structured, not something to structure with. In Z2 you had your
content objects, with rather big APIs, with realworld meaning, like a
page or document. They basically build the model. In z3 we try to
reduce that. It is no solution to say to the TTW developer here look
at all these interfaces and choose the one you need, configure them
with zcml and then start to use them in your script.
This supports in my view, that scripts are only useful as part of
something above core zope3.
4. Editing and testing without restarts.
5. As a question: Are scripts without acquisition useful?
Function I would not see are:
1. Editing through a webpage
2. funny parameter definition and context magic
With regards,
__Janko Hauser
_______________________________________________
Zope3-dev mailing list
Zope3-dev@zope.org
Unsub: http://mail.zope.org/mailman/options/zope3-dev/archive%40mail-archive.com
