[
Lists Home |
Date Index |
Thread Index
]
- From: Rick JELLIFFE <ricko@geotempo.com>
- Date: Thu, 12 Oct 2000 21:15:20 +0800
Dylan Walsh wrote:
> Perhaps there is a need for a formal or ad hoc standard for this reporting,
> e.g. an extension to SAX, that specifies subclasses of SAXParseException for
> different schema errors. Or is this already being dealt with?
There is something else that needs to be checked.
After processing a document using a schema, the information set is
augmented
by various items: type, etc. Different processors can augment the
infoset with different items, depending on what the application needs.
But there are three+ issues that need to be addressed AFAIK before there
can
be a SAX version of the PSVI (post-schema-validation infoset, doncha
love it)
* how to represent error and validity conditions (as Dylan mentioned)
* how to represent PSVI items that do not have any (possible) markup
equivalent: for example, the type of an element can be expressed
by adding an xsi:type attribute to the sax stream, but what can be
used to express the type of an attribute?
* does there need to be a special handling in SAX for element with
the xsi:null attribute (probably not)
So a post-schema-validation version of SAX may need to add arbitrary
properties on existing information items. I wonder what people think
the best approach for handling this might be?
Rick Jelliffe
|