[Zope3-dev] Protocol support was [Zope3-checkins] CVS: Zope3/lib/python/Zope/App/OFS/Services/SmtpService/Views - __init__.py:1.1 configure.zcml:1.1

[Gary Poster]

Tim's observations make a lot of sense to me.  Two opinions:

1) The core release of Zope 3, especially as conceived during this ramp
up to an actual release date, should concentrate only on a few
protocols, and we should look at any new protocol proposal with a hard
eye.

2) However, I propose we *should* have a "how to" written for initial
release (perhaps as part of Jim's tutorial) on how to build Zope 3
pluggable protocols.  Work on this should *replace* work on any extra
cool protocols for now.  This document should also be a significant way
for us to evaluate the architecture as we look for things that need
refactoring--that is, the instructions should describe an interface that
is not unnecessarily difficult to implement and connect, and if it is,
we should polish until it isn't.  

*If some people are willing to guide me or get me started on # 3 (who? 
I know Stephan knows this but is busy at Doctorate) then I'm willing to
write up a first draft.*

----

Does any of this seem reasonable?

Gary


On Mon, 2002-10-28 at 22:21, Tim Hoffman wrote:
> Hi Guido 
> 
> Actually I have yet to play with Twisted.
> 
> Actually my desire to utilise new protocols came from 
> some of the research I am doing (paid for, not personal) into 
> community frameworks. 
> 
> One of things we would like to achieve is provide many different 
> access channels into an online community, where the line
> between content access, messaging etc.. start to blur.
> 
> And zope as an application server rather than content server 
> actually provides a pretty good platform for playing around with this
> sort of stuff.
> 
> Personally I have thought that Zope could be a really good mud/moo
> platform. ie build the objects TTW, but interact through other protocols
> such as telnet/jabber ....
> 
> I suppose that all the talk of "not" making Zope3 CMF specific but
> rather it is an application server, that one can build CMF's on, 
> makes me think we shouldn't limit the scope of applications or
> for that matter how one can access them.
> 
> Regards
> 
> Tim