RE: [xml-dev] Stability of schemas -- frequency of versioning

[cbullard@hiwaay.net]

I'd have to say from the tool building perspective, I'm disappointed  
so far with what I am seeing of DTD support in system.xml from  
Microsoft.  OTOH, I'm just refamiliarizing myself with system.xml and  
I may have missed some important knowledge base articles.

There is a tendancy to think DTDs went the way of the dodo.  In MIL  
land, this simply isn't true.  DTD design is advanced and has produced  
workable by quite complex expressions.   One can chop these down to  
minceable nuggets (eg, at the work package level), but then getting  
these into validatible form and validating input coming from non-XML  
authoring tools (say Word) becomes a significant challenge.

IOW, I think of these questions in terms of the toolkits and  
frameworks, the notion being if an interface, human or otherwise, can  
get me documented well-formed XML, I can do the rest but I still have  
to live with the GFI.  Sometimes the harder problem is explaining that  
to the technical writers and their managers who still want to believe  
a well-styled manual is a good one.

len


Quoting Norm Birkett <Norm.Birkett@reval.com>:

> Which leads me to a question for the list: Have you observed any tendency
> among software developers to slight the design of XML languages, by
> comparison with other interfaces (such as database schemas or APIs)?
>
> If so, do you think that's because they suppose that tagged data is
> somehow easier to change at the schema level than non-tagged data?
>
> (I ask because I'm trying to get a sense of how software engineers tend
> to think about XML schemas, and you all have WAY more experience with this
> than I do.)
>
> Norm Birkett