RE: PIDF - using xml schema

At 04:40 PM 4/19/02 +0100, Adrian Bateman wrote:
>I've just finished reading draft-hollenbeck-ietf-xml-guidelines-00.txt
>and have the following observations for consideration:
>
>* draft-ietf-impp-cpim-pidf-02.txt:Section 3.1
>
>I think we should change this to REQUIRE an XML declaration. We should
>mandate the use of UTF-8 and recommend that the encoding attribute
>SHOULD be used to explicitly note the use of UTF-8.

I'm not up to speed on the detail issues here, but the XML declaration 
seems to me rather like excess baggage.  I have a (weak) preference to go 
the other way; i.e. prohibit it.  (xml-guidelines say to be explicit, 
IIRC.)  As for mandating UTF-8, I agree.

>We should also require that documents MUST be well-formed.

Yes.

>* draft-ietf-impp-cpim-pidf-02.txt:Section 4.1.7
>
>I think we should mandate the use of the xml:lang attribute for the
><note> element. Should it be possible to have more than one <note>
>element for different languages?

I think yes, on both counts.

>* draft-ietf-impp-cpim-pidf-02.txt:Section 4.4
>
>Rather than the DTD, we should use an XML schema and insist that PIDF
>documents MUST be valid according to the schema.

Again, I'm not up to speed on the subtleties.  But IIRC, XML schema handles 
the extensibility more smoothly so might be a better choice.  I think it is 
worth leaving the DTD in as a (non-normative) appendix because many folks 
are more familiar with that.  If nothing else, it will help these folks to 
understand the schema.

>On 19 April 2002 08:41, Graham Klyne wrote:
> > I'm not familiar enough with XML schema to check this in detail, but I
> > think it would be a good idea to include an XML schema as an appendix.
>Any
> > examples in the document should be validated against this schema.
>
>I had a quick look - at first glance the only thing that flagged up was
>using xs:dateTime for the timestamp. I'm not sure how this compares to
>the IMPP datetime format that we mandate.

Oh yes -- I think they're very close, but it's worth checking.  I suppose 
we ought to go with the XML spec here if there's a conflict.

#g
--

-------------------
Graham Klyne
<[email protected]>



  [reminder: [email protected] for non-technical discussions, please]