[dsdl-discuss] DSDL-9 - issues for discussion prior to / during WG1 conference call on August 6th

I am currently working on revising the text of DSDL Part 9 in accordance
with the Disposition of Commons on the CD text that was balloted earlier
this year - see SC34 Nxxx.

I have a couple of queries on which I would appreciate advice. Whether
or not these queries can be resolved in the next few days will determine
whether or not I am able to produce a new draft text for discussion at
the WG1 conference call next Sunday.

1. Unqualified attribute names

My understanding is that unqualified attribute names are not in any
namespace, according to the Namespaces in XML W3C Recommendation. In
other words, if a default namespace is declared, it only applies to
unqualified element names and not to unqualified attribute names. In
DSDL do we work within this constraint or not? If we do, we will need to
remove from DSDL Part 9 the current mechanism for binding unqualified
attribute names to the default namespace.

2. The DSDL-9 validation process

In attempting to deal with the principal Japanese comment, we need to
decide whether a DSDL-9 aware parser must be capable of reporting
/simultaneously /that an instance is invalid in legacy terms but valid
in DSDL-9 terms (or vice versa), according to whether or not DSDL-9
declarations that bind prefixes or names to namespaces are taken into
account.

I had assumed that, in order to simplify the development of DSDL-9 aware
parsers, it should be possible to transform a DTD or instance or both
prior to a normal parser validation process. However, such a
transformation would, without special provisions, make it impossible
subsequently to determine what the result of validation prior to
transformation would have been, so simultaneous validation in DSDL-9 and
legacy terms would not be possible. Special provisions could, for
example, include as part of the transformation process the addition of
fixed attributes to the DTD to indicate the state of names prior to
transformation.

So, should we insist that simultaneous validation in both legacy and
DSDL-9 terms is a requirement, or should we allow DSDL-9 parsers to be
silent as to the validity of an instance in legacy terms? Sequential
validation would always be possible using a pipeline validation process
using both legacy and DSDL-9 aware parsers. Insistence upon simultaneous
validation will make it necessary to embed DSDL-9 awareness more deeply
within the design of the parser.

Francis Cave
Editor of DSDL Part 9

--
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)