Re: XLink 1.1: Xlink vs "legacy" linking

/ Bjoern Hoehrmann <derhoermi@gmx.net> was heard to say:
| Let FooML be a XML format with legacy linking mechanisms that allows use
| of arbitrary foreign attributes such that a conflict between XLink 1.1
| linking and the legacy linking could arise. I want to make a FooML +
| XLink 1.1 implementation. No errors, same problem. Also note that XHTML
| M12N allows to make a XHTML 1.1 + XLink 1.1 document type.

Take any markup language, FooML, with elements that have some
semantics.

Take any markup language, BarML, with global attributes that have some
semantics.

Combine FooML+BarML.

If BarML claims that its semantics win then you lose interoperability.
FooML applications that don't understand BarML will do the wrong thing.

If BarML claims that its semantics lose then you lose interoperability.
BarML applications that don't understand FooML will do the wrong thing.

The only spec that can say what happens is FooML if it acknowledges
the possibility that BarML attributes will be used.

A "FooML+BarML" application written in the absence of a FooML spec
that gives guidance on this point is going to have to make a choice or
give users an option or something.

I simply don't think the BarML spec can hope to give the definitive
answer. All it can say is what the BarML semantics of such elements
are.

That's exactly what XLink 1.1 does.

                                        Be seeing you,
                                          norm

-- 
Norman.Walsh@Sun.COM / XML Standards Architect / Sun Microsystems, Inc.
NOTICE: This email message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution is prohibited.
If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.