Re: [wss] Withdrawing my issue (was RE: [wss-comment] Enumerations ofQName fault codes)

["Eve L. Maler" <Eve.Maler@Sun.COM>]

I think it was useful to bring this up anyway (and *I* owe *Irving* an 
apology, because we discussed this issue last week before he brought it 
to the attention of this group, and I completely missed that this was an 
enumeration of QNames at the time; clearly others did as well).

This unioning mechanism doesn't actually buy any more in the way of 
machine-readable constraints; allowing "1, 2, 3, and all the other 
positive integers" is no more powerful than allowing "all the positive 
integers", nor does it validate that the values in your own namespace 
are only the ones you defined.  And it makes the schema more subtle and 
complicated, which could work against the benefit of highlighting for 
the casual reader any values that the spec gives special semantics.  To 
boot, if you wanted to convey the special semantics of the called-out 
values in any machine-readable way (e.g., structured <documentation>), 
you have to go to quite a bit more trouble.  Thus, the spec is still 
required reading.

So, the unioning choice is not entirely inarguable.  It remains a neat 
trick with some benefits and some deficits.

However, given my track record in the past week, I'm willing to believe 
that my analysis misses about ten other relevant points. :-)

	Eve

Reid, Irving wrote:

> I owe the TC an apology. Martin Gudgin has pointed out the subtle but critical distinction that I missed: because the type is specified as an enumeration of QNames, there is no problem with "locking in" the wsse: prefix. This would only have happened if the type was specified as an enumeration of *strings*.
> 
> So, re-start the presses, move-along-nothing-to-see-here, and all the other clichés, and I'll go sit in the corner for a while.
> 
>  - irving -
-- 
Eve Maler                                        +1 781 442 3190
Sun Microsystems                            cell +1 781 354 9441
Web Products, Technologies, and Standards    eve.maler @ sun.com