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: Namespace: what's the correct usage?

Hi -

I usually just lurk on this list, and often don't have the time to follow
even one thread.  However, this particular discussion is developing
into War and Peace (plus Doctor Zhivago), and this weekend
I have had a little extra time to read...

Now I have an urge to throw in my two cents' and see if anyone
else concurs.

My thought is that we are encountering a classic conflict.  We would
like to develop a data interchange convention which would allow new
systems/processors to come online and begin communicating *without*
a priori knowledge.  We would like to enjoy a data interchange convention
which requires a minimum of markup (overhead).  We would like to
pack as much of the necessary information as possible into one
transaction, due to bandwidth, processing time, or what-have-you.  That
is, if namespace URIs are actually references to pages somewhere...we
don't want to be compelled to download those pages before processing
our XML data, do we?

Our "likes" are in conflict with each other.

We know there is no real magic here.  We must always have some form
of a priori knowledge, even if it is only that we know XML is coming to us
as a character stream (ASCII, utf-8, whatever).

If we want to ship all necessary knowledge in one transaction, and if we
live in the hope that unanticipated recipients will find our transmissions
and if we acknowledge that we have little control over the recipients'
interpretations,  then the best thing we can do is specify everything we
possibly can. That is, explicitly name *everything*.  Ship the DTD/schema
along with the data file...whatever it takes.

If our highest priority is to move data out, to impose as little overhead as
possible...if we know absolutely that the recipient will treat the content
as we intended...then we can do pretty much what we want.  Don't use
any namespaces!  Settle it all in side agreements, private ICDs, or as
we choose.  Of course, it's fragile, and it breaks.  And it's never a
resource for any system(s) other than the ones whose designers have
conferred with us.

Maybe I am not seeing all the subtleties here.  I did come into this
conversation partway through.  And it is late now, and I am tired.  But it
seem to me that there is no "one-size-fits-all" choice, ever.  It seems that
we each have to exercise judgement in making system tradeoffs.
It's good that there are people who care enough about these issues to
debate them far into the night, and that there's a place for those people
to do so.

Aura Sheffield
Software Developer

The xml-dev list is sponsored by XML.org, an initiative of OASIS

The list archives are at http://lists.xml.org/archives/xml-dev/

To unsubscribe from this elist send a message with the single word
"unsubscribe" in the body to: xml-dev-request@lists.xml.org