[
Lists Home |
Date Index |
Thread Index
]
Patrick,
But that is the rub here, is it more complicated?
Frankly schema works for tiny small schemas with few if any
logical branches to the structure - but even there you run
into difficulties - e.g. stock quote lookup, how tough is that?
Put in stock ticker symbol - get back quote. Hmmm, but then the
ticker is related to exchange, and status of the company, so
could have prefix with colon separator, and so on - suddenly
a tiny little something - and schema really cannot provide
enough validation already. Unless you add more fields, and have
prefix, exchange, ticker, name, but now you need rules - if I
enter name and exchange, but not prefix and ticker.... how can I
use schema to check the right permutations for me? But wait, I
can declare varients of just the permutations of structure that
I want and label them accordingly. Now I need a way for people
to know what context to use what structure construct. Oh - how
I do that? (hint - you need context variables).
Then as soon as you get into significant information exchanges
with dozens of inter-related fields - oh my!
Schema really does have a very limited use profile. That is the
way it was defined of course - they did start with a set of
requirements. Trouble is few if any people who are using schema
have bothered to read what the original design was intended to
do. 9 times out of 10 it ain't what people are now trying to
use schema for!
No surprises then that it is sub-optimal for their needs.
DW.
Quoting Patrick Yee <kcyee@cecid.hku.hk>:
> On top of Schema checking, an even upper layer may offer checking of
> semantics, too. This is, IMHO, a new concept which does not appear in
> the TCP/IP stack. Needless to say, it is more complicated than Schema
> checking.
> Regards, -Patrick
>
>
> Jimmy Zhang wrote:
>
> >IP packets are much simpler to check because the head
> >fields are fixed in length. CRC checking is typical,
> >and is equvalent to signature of SOAP.
> >
> >Other checking, such as IP, port, process id, TCP flow
> >control etc, are less sophiscated than Schema.
> >
> >Deep content inspection offers to look at teh packet
> >content,but still is less complicated than Schema.
> >
> >--- costello@mitre.org wrote:
> >
> >
> >
> >>Hi Folks,
> >>
> >>I remember reading a long time ago a posting by
> >>David Megginson. Paraphrasing, David said, "when
> >>dealing with XML you are working down at the bare
> >>metal".
> >>
> >>Other technologies work down at the bare metal, such
> >>as TCP/IP. Perhaps there are lessons to be learned
> >>there? Certain TCP/IP packets are rejected as bad
> >>and the other packets are accepted and passed up to
> >>other layers, where those layers perform additional
> >>constraint checking.
> >>
> >>If we accept the conjecture that XML is down at the
> >>bare metal level, then perhaps the only constraint
> >>checking that should be done by an XML Schema is
> >>parallel to the kinds of checking that is done on
> >>TCP/IP packets. All other constraint checking is
> >>passed up to other layers.
> >>
> >>Can someone characterize the kind of constraint
> >>checking that is done on TCP/IP packets? What would
> >>be the parallel constraint checking in XML Schemas?
> >>/Roger
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >-----------------------------------------------------------------
> >
> >
> >>The xml-dev list is sponsored by XML.org
> >><http://www.xml.org>, an
> >>initiative of OASIS <http://www.oasis-open.org>
> >>
> >>The list archives are at
> >>http://lists.xml.org/archives/xml-dev/
> >>
> >>To subscribe or unsubscribe from this list use the
> >>subscription
> >>manager:
> >><http://www.oasis-open.org/mlmanage/index.php>
> >>
> >>
> >>
> >>
> >
> >
> >-----------------------------------------------------------------
> >The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> >initiative of OASIS <http://www.oasis-open.org>
> >
> >The list archives are at http://lists.xml.org/archives/xml-dev/
> >
> >To subscribe or unsubscribe from this list use the subscription
> >manager: <http://www.oasis-open.org/mlmanage/index.php>
> >
> >
> >
>
>
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
>
> The list archives are at http://lists.xml.org/archives/xml-dev/
>
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://www.oasis-open.org/mlmanage/index.php>
>
>
|
