[Zope] Character-set woes (possibly OT)

[Mark Barratt]

I'm having trouble understanding how Zope accepts, processes and 
transfers character data - or more precisely what the character-set 
rules and pitfalls are in a workflow in which

. writers work in Word (or something)
. copy is pasted or typed into a Zope form field via a rich-text browser 
widget
. the field is filtered (to remove Word-HTML junk markup and characters 
we know might blow up MySQL) via a Python external method
. into (non-UTF) MySQL
. and back into a zope variable and redisplayed.

Presumably character-set use/'tolerance' is relevant in
WP application
Browser
Rich-text widget
Python
Zope
MySQL

...and their interactions?

Has anyone any experience or know of any information sources which might 
help me understand/control this workflow so that users have some chance 
of seeing what they thought they input?

Thanks

-- 
Mark Barratt
Text Matters
phone +44 (0)118 986 8313
fax +44 (0)118 931 3743
email markb@textmatters.com
web http://www.textmatters.com