Re: attribute order (RE: Syntax Sugar and XML information models)

["W. E. Perry" <wperry@fiduciary.com>]

This illustrates that the semantics of an XML instance may be substantially
different for different users, and that those semantic differences are
elaborated by the local processing applied for their own different purposes
by those users. In performing that processing, users have clues outside of
the canonical infoset of XML--in this case document order (and document
order of the DTD, if applicable). The needs of each user are as they are
and, given necessity, the user must find a means of local processing which
meets those needs while working with an XML instance as received. I believe
that this demonstrates how the creator or transmitter of an XML instance
should not presume what the semantic outcome of that instance will be in the
hands of a user and, more particularly, never expect that an XML
instance--whatever standards it might be created to--conveys specific
semantics. It is in the nature of XML that it will be processed anew by each
recipient. Given that the outcome of that processing must met the local
needs of each user, we must expect that processing to be locally
idiosyncratic.

Respectfully,

Walter Perry

Jeff Greif wrote:

> Should the person editing have to know that different orders have
> different meaning to the recipient?  What if different orders are
> necessary for different possible recipients?