Re: AW: ISO 8601 anyone?? And more on Parties.

[Jon Farmer <jfarmer@ic.net>]

If the issue is about "justifying redundancy" in representation of data in a
message, maybe it will help to consider the justifications of redundancy in
storage:

- performance because (since copies can be kept closer to heavy-reader
processes - "caching")

- availability (because if the source copy goes down the replicas may still
be up)

- queriability (because the copies can be of faster and more sensible and
flexible structure "normalization")

- manageability of security (when there are many reader users relative to
writer users, and there are many systems creating the content, it is far
cheaper to manage user ids and privileges around one storage rather than all
the soruces)

I think that in the case of messages in particular, these factors are valid
or invalid as follows:

- performance (redundancy int he message decreases the performance of
transmission but reduces processing load on receiver

- availability (redundancy in the message makes the equiavlent more
avialable - could accellerate uptake by vendor community)

- queriability (XML query tools would be directly usable for queries that
require uniformity of time format for time-based joins)

- manageability of security (actually I think it poses a security risk,
since senders could put in bogus eqiuvalents)

- storage capacitiy (increases required net capacity - so it is still a
cost)

- responsibility to keep it synchronized (in messaging the resposibility
lands on the sender but is recieve from the receiver)

I don;t think it is our place to preempt the ebXML community from optimizing
deployments in all these diomensions.  Therefore I would recommend that we
"allow but not mandate" multiple expressions of things (like times) but be
able to designate the "authoritative" instance in the set (such as the first
in the sequence).  We could also provide free of charge a "core component"
that provides for conversions.

Jon Farmer
----- Original Message -----
From: John McClure <hypergrove@olympus.net>
To: Ian Galpin <g1smd@freeuk.com>; ebXML core <ebxml-core@lists.ebxml.org>
Cc: Aron Roberts <aron@socrates.Berkeley.EDU>
Sent: Sunday, April 22, 2001 8:37 PM
Subject: RE: AW: ISO 8601 anyone?? And more on Parties.


> Ian says:
> > I am against repeating data in other formats. Use just one, and
> > one that is globally unambiguous.
>
> Fine, however, that is a different statement from saying you are against
> having an architecture for XML elements that accommodates multiple
> presentations, in multiple languages, of information stated otherwise in
an
> unambiguous "globally-standardized" format, eg 8601.
>
> It's ok to believe that, as a publisher, you wouldn't format a datastream
> with multiple presentations of a single piece of data. It's not ok to say
> that others may not do so. Or even should not do so.
>
>
>
> ------------------------------------------------------------------
> To unsubscribe from this elist send a message with the single word
> "unsubscribe" in the body to: ebxml-core-request@lists.ebxml.org
>