Re: [xml-dev] NVDL as a dispatching architecture

[Tatu Saloranta <cowtowncoder@yahoo.com>]

--- "MURATA Makoto (FAMILY Given)"
<eb2m-mrt@asahi-net.or.jp> wrote:

...
> However, the NVDL specification is silent about such
> application
> frameworks.  I believe that we need much more
> experiences with NVDL
> first.  Also, I am not sure that standardization is
> the right approach
> for designing such application frameworks.  (E.g., 
> SAX, DOM, and even
> STAX may be used for constructing them.)
> 
> I am hoping that early adaptors of NVDL accumulate
> board experiences 
> and make a hundred flowers blossom.

This seems reasonable.

It is interesting to read about NVDL, since
(for what it's worth) I have written a simple
validation framework that builds on StAX. This was
done as part of experimental Stax^2 set of extensions
on basic stax api, and allows for multiple validators
on individual input (parsing) and output
(serialization streams), including possibility to
start/stop individual validators mid-stream. I have
not added any support for inter-validator co-operation
at this point, although that would be a possibility
(within limitations of the streaming model).
Currently validators exist for basic DTD (provided
natively by the stax impl), and RNG (using MSV);
adding one for w3c schema is planned (using MSV as
well).

This approach is obviously less generic than NVDL, but
seems reasonable for efficient validation within
context of stax (pull) processing.

-+ Tatu +-


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com