Comment on 1.6 of 26300

[MURATA Makoto]

Dennis,

Thank you for your reply.

2013/7/21 Dennis E. Hamilton <dennis.hamilton at acm.org>:
> I'm afraid I wrote that.
>
> Yes, ODF Consumer would be correct.  I don't know why ODF Processer is there.  Whatever the reason, it is apparently a bad idea.  This is for ODF 1.0/1.1 so Conformance Targets are not so explicit as in ODF 1.2.  Consequently, saying ODF Consumer doesn't help, I believe, even though it is correct (and used in ODF 1.2 Part 1 section 3.18).


>
> The best term would be simply "processor."

What does it mean?  Is it different from XML processors, which a
are clearly defined in  XML?

>
> The reference to the RNG Data Model is intentional.
>
> The idea was to find a place in the RNG Data Model where an element child entirely of white-space characters that is neither element text nor part of a data type lexical form could always be ignored completely.  I don't recall whether this was also intended to capture any case where a regular expression might also apply.  The term pattern was meant in the sense of Section 9.3 of ISO/IEC 19747-2:2003.  If it is already the case that such white space would be ignored in accordance with RNG rules, the statement in ODF 1.0/1.1/1.2 can be simplified.

I am reasonably knowledgeable about RELAX NG and
I think that the different between "weak match" and "match"
is one of the hardest parts of RELAX NG.  Are you trying to
eliminate whitespace text chunks in the child sequence
that weekly match but does not match the pattern
so that the sequence match the pattern?  I think that
this is too complicated.  I think that we do not have
to force ODF applications to ignore such text chunks.


>
> There may also be an interaction with the rules for eliminating or unwinding foreign elements in places where the foreign element occurs in a paragraph-content sequence.  In this case, the consumer might not have any knowledge of the schema for the foreign element, yet there are rules for preserving its content.  I do not think that was handled.

I strongly think that we should delete this
subclause.

Regards,
Makoto
>
> -----Original Message-----
> From: sc34wg6-bounces at vse.cz [mailto:sc34wg6-bounces at vse.cz] On Behalf Of MURATA Makoto
> Sent: Saturday, July 20, 2013 02:42 AM
> To: SC 34/WG 6 mailing list
> Subject: Comment on 1.6 of 26300
>
> Since reviewers of the ODF JIS are tough, I need
> help from you guys.
>
> 1.6 of 26300 (as corrected and amended by CORs and an AMD),
> has a para shown below:
>
> In addition, ODF processors shall ignore all element children ([RNG]
> section 5, Data Model) of ODF-defined elements that are strings
> consisting entirely of whitespace characters and which do
> not satisfy a pattern of the ODF schema definition for the element.
>
> First, what is an "ODF processor"?  It is never defnied.
>
> Second, what is "element children" as defined in RELAX NG?  The
> only term I can find is "an  ordered  sequence  of  zero  or
> more  children;  each  child  is  either  an  element  or  a
> non-empty  string;  the sequence never contains two consecutive
> strings".
>
> Third, but technically most importantly, "do not satisfy a pattern" is
> at least misleadnig.  If it is reworded as "do not match a pattern",
> it would be technically correct.   But it continues to be very
> misleading, since readers are required to tell the difference between
> "week match" and "match" in RELAX NG.  Wait.  We can
> consider matching only for a sequence.  Not for a string in a
> sequence.
>
> Regards,
> Makoto
> _______________________________________________
> sc34wg6 mailing list
> sc34wg6 at vse.cz
> http://mailman.vse.cz/mailman/listinfo/sc34wg6
>



-- 

Praying for the victims of the Japan Tohoku earthquake

Makoto