[dsdl-discuss] Re: Selection and Dispatching are different

Alex Brown wrote:

>Rick hi
>
>
>
>>The discipline
>>of "XML-in, XML-out" means that any mechanism for transmitting the XML
>>infoset can be used
>>
>>
>
>Aha, so when we say the input to a validator is well-formed XML, we don't
>necessarily mean that the data being validated is itself well-formed, merely that its XML infoset representation is ?
>
>
An infoset is not a representation. It is a set of the types of
information in a document
that is of interest for some application. We limit ourselves to
well-formed (parsed) XML
documents to enforce a discipline. Because by doing so we reduce the
amount we need
to define, we follow the "principle of least surprise", we get DSDL out
within our
limited resources, we maximize the applicability of all the generic XML
technologies
and libraries and so help implementers, and we reduce DSDL to an n-1-n
kind of
architecture (xml/sgml/etc to XML to validators) which gives a hope of
more formal modelling
and theoretical characterization. (Also, by not defining too much in
terms of an abstraction,
we prevent unserializability and allow loose coupling if that is needed.)

I should point out that my company has a product out this week that does
Schematron
validation of SGML. We may add RELAX NG etc. We just convert SGML to ESIS
with SP, then I wrote an ESIS to SAX converter. But conceptually this is an
XML-in system as far as validation is concerned: only XML
information-set information
makes its way to SAX. We get line locators fine, for error tracking.

So just because we talk of XML doesn't mean that SGML is out in the cold:
just SGML information that does not have an equivalent in XML
information-set
information.

Cheers
Rick Jelliffe

--
DSDL members discussion list
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-discuss-request@dsdl.org
(mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)