Re: [Xerces2] Clarification request: item 22 The parser shall support grammar access for both DTD and Schema.

[Andy Clark <an...@apache.org>]

Ed Staub wrote:
> What does "grammar access" mean?

Grammar access means different things to different people so
this is definitely something that we need to lock down in order
to understand what is required.

> Does it include an API which directly provides editors with 
> an answer to "what can go here"?

This is clearly something people want to do. In addition, they
want to validate a document based on the grammar of their
choosing. Then you have the problem of choosing what kind of
access to the document is required for the standard validation
mechanism. SAX? DOM? Or do we just provide some implementations
that know how to validate SAX streams and DOM trees against the
specified grammar(s)?

> Does it include an API for walking the grammar at all?

Both query and editing of the grammar will be requested 
features. How much do we support, though?

At the bare minimum, the internal parser must have enough
information to populate standard grammar access APIs. For
example: SAX 2 callbacks and DOM Level 3 content models.

The design of the concept implementation can answer some
of these questions but not all of them. The focus has been
more on validating instance documents in a consistent way
using a universal validator and grammar objects. So this
is an area that needs to be explored more.

-- 
Andy Clark * IBM, JTC - Silicon Valley * andyc@apache.org