RE: Restricted Character Sets

Hello Hitoshi,

We really appreciate your review.

> In 7.1.10.1 Restricted Character Sets,
> it is better that this feature allows optional
> because it takes much time to compute the restricted
> characters based on pattern facets of a schema beforehand.

The working group had internal discussions before, regarding this
issue. At that time we came to the conclusion and still believe that
computing restricted characters does not pose a computational
issue. There are essentially two reasons for our thinking.

First of all, computing char sets in EXI will be done once statically
when parsing an XML schema. While coding an XML instance to EXI or
vice-versa no additional computing is required.

Furthermore, we are in receipt of an implementation report shared by
working group members which supports the assertion that computing the
char sets is not an overly complex process. XML validators often use 
XML schemas to do similar character set calculation based on patterns, 
which involves even more computation than EXI requires. Yet they are 
able to process the schema in a reasonable amount of time.

We are cautious about making it possible to turn on/off features
arbitrarily since doing so repeatedly would accumulate and tangibly
affect the compactness and processing efficiencies. So we would need
convincing arguments before making this sort of change.  We are 
interested in hearing more about the basis of your concerns.

Thanks,

--mike


Mike Cokus (for the EXI Working Group)
The MITRE Corporation
757-896-8553; 757-826-8316 (fax)
903 Enterprise Parkway
Hampton, VA 23666

>-----Original Message-----
>From: public-exi-comments-request@w3.org [mailto:public-exi-comments-request@w3.org] On Behalf Of >UCHIDA Hitoshi
>Sent: Thursday, November 06, 2008 5:59 AM
>To: public-exi-comments@w3.org
>Cc: youenn.fablet@crf.canon.fr; fujisawa.jun@canon.co.jp
>Subject: Restricted Character Sets
>
>
>Dear EXI WG,
>
>In 7.1.10.1 Restricted Character Sets,
>it is better that this feature allows optional
>because it takes much time to compute the restricted
>characters based on pattern facets of a schema beforehand.
>
>--
>Regards,
>Hitoshi Uchida
><uchida.hitoshi@canon.co.jp>

Received on Friday, 25 September 2009 14:18:14 UTC