- From: John Schneider <john.schneider@agiledelta.com>
- Date: Wed, 14 May 2008 10:15:25 -0700
- To: <santhana@huawei.com>, <public-exi@w3.org>
- Message-ID: <03fd01c8b5e6$19f9dc90$6637a8c0@jcsdell8600>
Hi Santhanakrishnan,
Yes, you are right. This is an intentional optimization to increase
processing efficiency and simplify implementation. Pruning the unused CH
events that have length 2 would reduce processing efficiency and complicate
implementations for only a small improvement in compactness (one bit for
some elements).
All the best,
John
_____
From: public-exi-request@w3.org [mailto:public-exi-request@w3.org] On Behalf
Of santhanakrishnan
Sent: Wednesday, May 07, 2008 2:10 AM
To: public-exi@w3.org
Subject: [EXI] CH Event processing
Hi
During encoding when we encounter a CH EVENT and if the current
grammar does not have a production with event code of length 1 matching CH
event we add a new production for CH with event code 0. This is quite
useful.
But why should not the existing CH production be removed from the
grammar. I feel it is simply a place holder and in some cases can lead to
extra bit for encoding than necessary(when there are 2**n productions).
So I suggest to remove the existing CH production when we are adding
a new CH production with event code length of 0.
Regards
Santhanakrishnan
Received on Wednesday, 14 May 2008 17:16:10 UTC