RE: [xml-dev] W3C's five new XQuery/Xpath2 working drafts - Still missing Updates

[Jonathan Robie <jonathan.robie@softwareag.com>]

At 02:24 PM 1/2/2002 +0000, Sean McGrath wrote:

>>[Michael Champion]
>>The larger issues that Mike Kay raises are critical: All this committee work
>>is for nothing if the result is too complex or expensive to actually use.  I
>>am not all that much dumber than the average software developer, I have
>>followed the XML world full-time for 5 years now, and this
>>schema/PSVI/strongly-typed XQuery stuff makes my head spin.  I can't imagine
>>what ordinary developers who don't focus on XML will think of it.
>>
>>Actually, come to think of it, I can ... it will be C++ and the Windows API
>>all over again; few developers go anywhere near it without GUI tools and
>>wizards to hide the complexity behind a proprietary front end.

Formalisms are a helpful way to design something that is simple and 
consistent. For most users, formalisms are not easy to grasp, so they are 
not a good way to present a language to a user. Languages quite commonly 
are designed using tools that the user is not exposed to - relatively few 
users read the EBNF description of the languages they use, learn the formal 
type system, or understand precisely how polymorphism is implemented. In a 
well-designed language, this does not prevent them from using the syntax, 
the types, polymorphism, etc based on an informal understanding.

RELAX-NG is a simple and straightforward language, but it is defined with 
formalisms quite similar to those of the XML Query Formal Semantics. XML 
Schema is not as simple and straightforward as RELAX-NG, and this may be 
partly due to the fact that its formal semantics were defined after the 
language itself - had it been done earlier, I think XML Schema would be 
simpler for end users.

I think XQuery is pretty easy and straightforward for end users.

>Thus playing right into the hands of those who would make lock us in
>to their tools. As data owners, we need to fight this. When really smart
>people start talking about needing GUIs to grok what some notation really
>*means*, they are wittingly or otherwise heading down the vendor-lockin
>fork in the road.

I disagree - the formalisms ensure that the language is unambiguously 
specified, which ensures interoperability.

Most programmers work perfectly well with a somewhat vague understanding of 
what their languages do - adequate for programming, but not adequate for 
implementing the language. I doubt very much that people will learn XQuery 
by reading the Formal Semantics. The language specification itself is also 
becoming more oriented toward implementors.

>I fully expect to get flamed for that statement but what the heck.
>I'm trying to grok some vendor locked in XML at the moment and consequently
>I'm not in a very good humour.

XQuery is being widely implemented - surely you are not claiming that it is 
a proprietary vendor scheme? If so, which vendor are you accusing? Since 
the original language, Quilt, was created by people from Software AG, IBM, 
and Crossgain, I assume one or more of these companies must have hatched 
the plot....

Jonathan