• From: "Michael Kay" <mike@s...>
• To: <noah_mendelsohn@u...>
• Date: Thu, 12 Oct 2006 00:08:02 +0100

> Yes, I think that's quite nice, but I don't think it quite 
> emphasizes what we wanted to convey for xsi:schemaLocation in 
> particular, which is to emphasize that there may be good 
> reasons why the processor won't even try to follow the 
> pointer. 

For xsi:schemaLocation I think the situation is quite different. Someone
receiving a document has the choice of what to do with it:

(a) don't validate it at all

(b) validate it against schema X

(c) validate it against the schema identified in its own xsi:schemaLocation

If they ask for (c), then it would be surprising if the processor then
ignored the value of xsi:schemaLocation and used some completely different
schema.

So in this case too, I think the "hint" terminology is unhelpful. It would
be better to say that the algorithm used to dereference a URI may be under
user control.

Michael Kay
http://www.saxonica.com/