[Zope-CMF] Re: A better CMFworkflow - XML configuration

Jeff Kowalczyk jtk@adelphia.net
Sat, 7 Jul 2001 02:24:47 -0400


[Shane] I've written several XML (de)serialization routines (for other
projects), and they all turned out to be a lot more work than I
expected.  There must be a pattern out there that's easy to apply and
extend.

[jtk] This type of thing needs to happen at the framework level to be
useful. XML (de)serialization is important because it always gives you a
point of disembarkment from your framework if it can't do, or takes
undue effort to do, a particular kind of data manipulation. In our case,
the highly-configured workflow tool could benefit from an alternative to
property pages, for reasons including, but not limited, to
visualization.

There's two general-case implementations that come to mind. Of course,
neither of them are for Python, but could inspire an implementation
strategy for Zope.

Java: http://www.csse.monash.edu.au/~bren/JSX/

.NET:
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpguidn
f/html/cpconserialization.asp 
(if one can browse msdn without IE, I'll be shocked)

As if compliments to MS would fly in *this* group, but the .NET class
libraries are impressive. Never mind about the C#/VB languages, it's the
Libraries that do the work. Look at what they can do with XML
serialization:
http://www.gotdotnet.com/quickstart/howto/doc/xmlserialization/RWObjFrom
XML.aspx

Turn something like *that* loose with ZPT, and look out!

I'm watching the Zope-NewReligon effort with interest, because to
preserve Python's productivity edge, we'd be better off imitating .NET
rather than IUnknown-era COM. They don't really query interfaces
anymore, they cast to interfaces and data types safely. This class
framework is going to give Zope some serious competition if they ever
get a ODBMS or XMLDBMS to manage what their classes can already do with
persistency.