Comment for EXI Draft - Grammar learning

Dear W3C EXI WG members,

I have a comment for this draft specification.

EXI has a mechanism of the grammer leaning (ref. Section 8).
So, an EXI parser use large memory to keep grammar data, if the XML data
has many kinds of node (ex. element).
For small devices, this is a serious problem.
I think the limitation mechanism of the grammar learning is needed in
the specification (ex. MAX number of the kind of event code).

Regards,
Keisuke Tamiya (tamiya.keisuke@canon.co.jp)

Received on Friday, 7 November 2008 08:24:52 UTC