[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: How to get the 1st element in DTDD
- From: Murali Mani <mani@CS.UCLA.EDU>
- To: XML Everywhere <firstname.lastname@example.org>
- Date: Fri, 22 Jun 2001 14:34:48 -0700 (PDT)
My scope is quite limited I believe -- I am not sure how
SOAP/Biztalk/ebXML etc work together with xml schemas ... -- all I do is
simple data modeling using xml schemas, some querying and some type
inferencing... that itself has been quite difficult :)
I want a schema to specify what should be the possible set of roots of
that document -- I actually see no reason why this should actually
eliminate its possibility to embed in another application -- you can
probably design a specific protocol for such embedding, if that is the
worst case -- but i am just saying the above ... i have no clue ...
but i believe in this -- a tree starts with a root, the root of an XML
document when you consider DB applications is often times "fictitious" in
the sense that it actually might be some type which is not present in the
database or something like that... but still, when you want a tree, I
think you should specify possible root elements ...
I am not sure if I should say this, but --
A regular tree grammar is specified as a 4-tuple (N, T, P, S)
N is the set of non-terminal symbols, which correspond to types.
T is the set of terminal symbols, which correspond to the actual element
tag names which occur in the document.
S is the set of start symbols, which is a subset of N
P is the set of production rules of the form
A -> x (R)
where A is a non-terminal symbol, x is a terminal symbol, and R is a
regular expression over non-terminal symbols.
The rule specifies that the non-terminal symbol A produces a tree with
root x and child nodes which correspond to R.
The above very simple specification, is actually very powerful -- actually
it is much *more* expressive than XML-Schema, DSD etc, in that you can
write a regular tree grammar for a set of trees, that cannot be specified
using XML-Schema etc.
Also the above regular tree grammars gives me the closure properties i
need -- closure under union and other operations (including query
operations), which unfortunately XML-Schema does not give.
I think the above abstraction works the best for me.
cheers - murali.
On Fri, 22 Jun 2001, XML Everywhere wrote:
> How about this.. once you specify root elements
> you preclude these root elements from being embedded
> inside other documents, such as a
> SOAP/Biztalk/ebXML envelope. I think
> all documents should be embeddable,
> and I can't think of any reason why you
> would not want to do that, except perhaps
> for security reasons (really stretching it).
> Or am I not understanding what you mean?
> ----- Original Message -----
> From: "Murali Mani" <mani@CS.UCLA.EDU>
> To: "Bing Li" <Bing.Li@asu.edu>
> Cc: <email@example.com>
> Sent: Friday, June 22, 2001 11:19 AM
> Subject: Re: How to get the 1st element in DTDD
> I have always believed that the schema should specify the
> possible root types, people have always said against this, I have always
> waited for one good reason for not specifying the root of the document in
> the schema, but unfortunately, I do not think I have seen any... I will
> keep waiting for some good reason ...