Re: Placement of the oasis-xml-catalog PI

[Paul Grosso <pgrosso@arbortext.com>]

At 13:54 2001 05 07 -0700, Lauren Wood wrote:
>Norman Walsh wrote:
>> / John Cowan <jcowan@reutershealth.com> was heard to say:
>> | Norman Walsh wrote:
>> | 
>> | > Now I think I do. I've looked over the SAX API again and I don't see
>> | > how the processor can tell when the <!DOCTYPE declaration has been
>> | > encountered...
>> | 
>> | Well, the first time that resolveEntity is invoked (if it ever is),
>> | the DOCTYPE declaration has definitely been reached.  So if the
>> | PI always comes before the DOCTYPE declaration, then it is guaranteed
>> | to have been processed before the first need to resolve anything.
>> | 
>> | IIRC we actually require the PI to come before anything that
>> | needs catalog resolution, such as a stylesheet link.
>> 
>> I'd like to express that requirement differently: the PI has no effect
>> on the resolution of external identifiers or URIs that occur before
>> the PI.
>
>I think this works; are there any oddities with internal and external 
>subsets that might trip us up? I can't think of any but just in case 
>somebody else can...


Wait, is Norm suggesting that the PI can even be within the internal
or external subset?  I'm certainly against it being in the external
subset, and I'm arguing for putting it up at the very top.