Re: Can XML Schemas Support Document Systems (WAS RE: ZDNet Schema article,and hiding complexity within user-friendly products)

[Murali Mani <mani@CS.UCLA.EDU>]

First of all, i personally think most of the stuff are not hard to
understand -- if they are hard to understand, then we can very well
suspect their correctness.

XML Schema might be able to support document systems, but I have not seen
a single operation defined for xml schema.

let me give one simple example -- consider a system that does the
following -- it performs some select operations on the documents, returns
the result document and also the result schema. I think such a system is
not definable at this point of time using XML schema.

to elucidate a bit further -- what can happen is the output schema can be
something like
a) (book*, author*, book*),
b) (book1*, book2*) etc
both of which cannot be defined using xml schema -- note that the first is
what "we" call horizontal ambiguity/non-determinism, and the latter is
what we call vertical ambiguity/non-determinism. If you want to get into
the "gory" details, i might recommend one of my own work -- "Reasoning
about XML Schema languages using formal language theory" -- it is
published as an IBM Technical Report -- available at
http://www.cs.ucla.edu/~dongwon/paper -- personally I think almost all
aspects of it is *quite* important.

When we start off with some implementations of such systems, we might
think it works -- but suddenly we get bad examples, and "unexpected things
happen" -- if we did not recognize this non-closure and somehow handle
this.

once again some history and also this might seem slight digression -- but
regular tree grammars and local tree grammars have been *widely* studied
for a very long time (local tree grammars were definitely studied in 75).
they are very clean mathematically -- local tree grammars are not closed
under several operations, but we can easily define most operations --
Makoto Murata has studied them. XML Schema, i am not sure how we are going
to go about defining operations. First thing probably is define union for
XML schemas. I might take a hint from history that local tree and regular
tree grammars are definitely mathematically precise and more interesting
than the types of grammars defined by XML Schema -- which does not conform
to either.

regards - murali.

On Wed, 25 Apr 2001, Bullard, Claude L (Len) wrote:

> On the other hand, I am interested in the assertions
> that XML Schema cannot support document-centric applications.
> It appears to me that it can if the right set of features
> are used, but maybe I am missing something more vital than
> "it's hard to understand".
>
> Can anyone summarize the position that it can't support
> these systems?
>
> Len
> http://www.mp3.com/LenBullard
>
> Ekam sat.h, Vipraah bahudhaa vadanti.
> Daamyata. Datta. Dayadhvam.h
>
> ------------------------------------------------------------------
> The xml-dev list is sponsored by XML.org, an initiative of OASIS
> <http://www.oasis-open.org>
>
> 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
>