OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Element Order significance in XML

> -----Original Message-----
> From: John Cowan [mailto:jcowan@reutershealth.com]:
> My advice: Learn to live with fixed order.  The additional flexibility
> is rarely worthwhile.

Ah, jeez, I hope I'm not starting another long thread, but:

When dealing with data sets, it would seem to me that order is
insignificant. Why? Because I have the same information if it's {A, B, C} or
{B, A, C} assuming there are no relationships between {A, B}. Why force me
to assume that their are? Even in XML Schema, there a limitations placed on
the <all> model group, so that it can be used nested in a <sequence>, for

So it seems like in most cases you're stuck declaring order significance
when in fact their may be none. IMO (uninformed and conjectural), the reason
this dependency exists is because the ML's were originally designed for
markup of documents that contained narrative, which is order-dependent. But
with XML being used for data transport, maintaining nonsensical sequence
information in the internal data model of the consuming application imposes
unnecessary overhead. 

What if I want to stick keyed siblings in a hashtable, for instance? Or what
if I want to build content by popping things onto a stack? Can't do it
naively, because things have to come out or be built in the right order to
be valid, even if that order has no significance. So you have to maintain
sequence information which is an artifact of the schema you're using. How is
that worthwhile?

Okay, flame on. I go wore my NOMEX pants today.

-- Jeff