[ubl-ndrsc] Fragment processing with local elements

["Eve L. Maler" <eve.maler@sun.com>]

Per my action item, I have asked some colleagues about difficulties in 
fragment processing.  If I get any additional input I'll forward it along.

	Eve

From: Norman Walsh <Norman.Walsh@Sun.COM>

/ "Eve L. Maler" <eve.maler@sun.com> was heard to say:
| Commerce One, having done some experiments around this, expressed the
| concern that if you break up your XML document into fragments and try
| to process those fragments independently, local elements at the
| "roots" of these fragments need to have their type specified with
| xsi:type in order to be schema-processable.

Yep. And if you don't have globally declared types for them, I don't
know what you do.

| That seems like a
| reasonable situation to me; a local element is only known within the
| scope of its parent so you have to provide more context.  But are
| there other gotchas?  Should fragment processing be done in some other
| way for better efficiency (i.e., why is the literal XML document being
| broken up into literal fragments on disk rather than a parse tree
| being manipulated)?

Short of passing around validated infosets (and can you really say
it's still valid if you rip a hunk out of the middle and pass it to
someone else?), I don't think passing trees around is much different
than passing serialized characters around.

                                         Be seeing you,
                                           norm

- --
Norman.Walsh@Sun.COM    | The public! The public! How many fools does
XML Standards Architect | it take to make up a public?--Chamfort
Web Tech. and Standards |
Sun Microsystems, Inc.  |
-- 
Eve Maler                                        +1 781 442 3190
Sun Microsystems                            cell +1 781 354 9441
Web Technologies and Standards               eve.maler @ sun.com