OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] XPath/XSLT 2.0 concerns

[ Lists Home | Date Index | Thread Index ]

Elliotte Rusty Harold wrote:
> At 2:23 PM +0200 10/4/02, Robin Berjon wrote:
>> In the case I am considering however, the document can only be valid 
>> because it has been binary encoded using a schema, and it won't encode 
>> if it ain't valid.
> OK. That may make sense for you, but it seems a very weird and unusual 
> use-case.

Not all that weird and unusual as it's already flying well, but imho not 
common enough to justify inflicting typed XPath/XSLT upon people that 
just need to query and convert documents. They'll just keep using the 
features they use today, only they'll be using bloated software to do 
it, ie pretty much the situation with Office. That's not desirable.

> If this is accurate, I don't think the cost of complexifying 
> the XPath 2.0/XQuery specs with schema awareness vastly outweighs the 
> benefit to strange applications that want to binary encode documents 
> according to a schema.

That's precisely (one of) the ideas behind this thread: we're looking 
for use cases that are solid enough to justify the added complexity. On 
this my position hasn't changed: we can appear to find potentially 
interesting use cases, but none that are of widespread applicability.

Even with widespread binary XML as described above, users will most 
likely rarely need the features of XPath 2.0. Some developers will, and 
that's why it can be done in a separate module or layer.

I see no good reason to not have XPath 2.0 Basic and XPath 2.0 
TypedWithExtraCheese (as the latter can be argued to be useful at least 
within XQuery context, and probably within others). Given that the XPath 
2.0 syntax is still in flux, the type-related extensions may even be 
specified to be in a different namespace (provided they're function 
based, or that axes can be namespaced).

> The implementation experience does not yet 
> exists to decided whether this is useful or even possible. It's better 
> to let this stuff be experimented with before anybody attempts to 
> standardize it.

Agreed. Hopefully, we're far enough from Rec that it may still be 
possible to provide feedback either way.

Robin Berjon <robin.berjon@expway.fr>
Research Engineer, Expway
7FC0 6F5F D864 EFB8 08CE  8E74 58E6 D5DB 4889 2488


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS