W3C home > Mailing lists > Public > public-rif-wg@w3.org > July 2008

Re: one thing we forgot

From: Michael Kifer <kifer@cs.sunysb.edu>
Date: Sat, 5 Jul 2008 13:49:57 -0400
To: Sandro Hawke <sandro@w3.org>
Cc: "Boley, Harold" <Harold.Boley@nrc-cnrc.gc.ca>, "RIF WG" <public-rif-wg@w3.org>
Message-ID: <20080705134957.00367165@kiferserv>


On Sat, 05 Jul 2008 13:31:30 -0400
Sandro Hawke <sandro@w3.org> wrote:

> > In the following I assume we do *not* differentiate dialects by pointing
> > from instance documents to the locations where their schemas are defined
> > (via xsi:schemaLocation, xsi:noNamespaceSchemaLocation, or similar:
> > http://www.w3.org/TR/2006/WD-xmlschema11-1-20060831/#xsi_schemaLocation
> > http://www.w3.org/TR/2006/WD-xmlschema11-1-20060831/#schema-loc).
> > 
> > What about having an ***optional*** attribute, dialect, on <Document>?
> > It optionally specifies the dialect the sender intends this Document to
> > be validated against.
> 
> I don't think that works.  The sender says it's Dialect 1, which the
> receiver has never heard of.  It happens to fall into the intersection
> of Dialect 2 (which the receiver implements, and the sender has never
> heard of) and Dialect 1, so it's perfectly usable.  But if the attribute
> is present, the document will have to be discarded, right?

What about "dialects" (as the attr name) instead of "dialect"?
The publisher would be able to specify several dialects
(with the or-semantics). The publisher is interested to make his doc as widely
available as possible, so he'll specify either the best known dialect (e.g.,
BLD) or as many dialects as is reasonable.

If the publisher specifies some rare dialect that nobody knows then, perhaps,
he does not want others to use that document - so it is ok.


	--michael  


> > The receiver is still free to reuse the Document in
> > novel dialects by trying to analyze its syntactic features.
> 
> Ah, so the attribute has no meaning or affect on the receiver?  Then
> what's the point?  Who is going to look at?  Only incorrect
> implementations.  Why do want to help them do the wrong thing?
> 
> > Isn't the official set of names, {BLD, DTB, PRD, ...}, of collections of
> > those features (Standard Extensions), and their intended use, something
> > RIF has to carefully keep track of anyway?
> 
> This matters mostly for non-standard (vendor) extensions, I think.
> 
>       -- Sandro
> 
> 
Received on Saturday, 5 July 2008 17:50:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:50 GMT