Re: I18N #5 -- proposal and QUESTION (corrected)

WG

Does no one have an opinion on the questions I asked, esp. #2?

Synopsis:  Application of the BLANK-HYPHEN-UNDERSCORE normalization of 
'cgmFont' and the font names in the metafile is unconditional in 2.1 
LCWD.  Should there instead be an attribute to control its 
application?  (E.g., it could have values apply/suppress, with default 
'apply' like now.)

Explanation:  I believe, as I wrote in I18N-5 proposed response, that this 
WebCGM-specific normalization is principally targeted the Adobe-13 core set 
of font names -- to normalize both current and legacy usage, in both WebCGM 
metafiles and other "close" profile metafiles.  Our explanation of its 
applicability was unclear.  I proposed a clarification (below attached -- 
do you agree with that?).

When clarified, if you agree with the proposed clarification it seems to 
invite the question, "Why apply that normalization unconditionally in those 
cases -- e.g., Asian font names encoded with Asian pages of UTF-16 --where 
it is unlikely to be useful?

I cannot see cases where its application would be harmful, but it seems 
kinda' dumb to unconditionally mandate its application to strange font 
names in some non-Latin languages, where we know it is inapplicable.

In other words, this is probably not a critical defect, but an oddity that 
emerges when one clearly understands the purpose of that 
normalization.  (Unless I am misunderstanding and misrepresenting the latter.)

-Lofton.

At 10:53 AM 12/16/2008 -0700, Lofton Henderson wrote:

>[...slight correction to the sentence beginning "I18N Comment #5 pertains..."]
>
>Hi WebCGM WG --
>
>Please see my questions, after "Proposed response"...the latter is what we 
>roughly discussed and agreed to in telecons.
>
>This is one of a series about the six I18N comments.  Although I18N sent a 
>separate message for each of their comments -- which is how we should 
>reply to I18N -- all comments are conveniently found in a single table 
>[1].  All comments apply to 9.3.2.2, the ACI mapList element [2]
>
>[1] http://www.w3.org/International/reviews/0811-webcgm/
>[2] 
>http://www.w3.org/TR/2008/WD-webcgm21-20080917/WebCGM21-Config.html#ACI-maplist
>
>In this message, I have proposed a response.  Email discussion 
>welcome.  We will resolve definitely in a WG teleconference ("as proposed" 
>if no discussion).
>
>I18N Comment #5 pertains to normalization of font names from the metafile 
>and the 'cgmFont' value.  It is labelled Substantive by I18N.
>
>Comment #5
>=====
>"These normalization rules are applicable for font names specified using 
>the characters of ISOLatin1. They will likely be inapplicable for font 
>names specified using other non-Latin characters."
>
>What happens in the case of Latin-2 (Eastern Europe), which is similar to 
>Latin1 but contains a few additional characters. Does a single non-Latin1 
>character cause normalization to be abandoned for the whole string?
>
>It seems like the only thing that wouldn't apply to all non-Latin1 font 
>names is converting to lower-case, though that is still a relevant 
>consideration for many other Latin characters outside Latin1, and for 
>Armenian, Greek and Cyrillic. Why restrict to Latin1?
>
>Proposed response:
>----------
>The apparent restriction to Latin 1 was unintended.  As you point out, the 
>normalization would work nearly the same if the same names were expressed 
>in Latin 2.  Latin 1 got the special mention because:  the default 
>character encoding of WebCGM is ISO 8859-1; and the vast majority of 
>current and legacy WebCGM instances use this character encoding and a 
>restricted core set of thirteen specific font names.  As pointed out in 
>WebCGM's reply to I18N's issue #3 (@@link@@), these WebCGM-specific 
>normalization rules were targeted at the substantial legacy and current 
>metafile volume that intend to invoke this restricted core set of fonts, 
>but that contain well-known, trivial deviations in the construction of the 
>names.  In other words, the real target is trivially deviant usage of the 
>13 specific core-font names, regardless of the character encoding.  (More 
>background:  the valid character encoding for any particular WebCGM 
>instance is one of the three ISO 8859-1, unicode UTF-8, or unicode UTF-16.
>
>WebCGM will reword to clarify the useful scope of these normalization 
>rules, to remove the implication of a normative restriction of 
>applicability, and instead to be advisory about the usefulness of that 
>normalization outside of its primary intended scope.  Replace the sentence 
>in question with:
>
>"Note:  These normalization rules are derived from and intended for a 
>substantial volume of existing metafiles that aim to invoke fonts from 
>WebCGM's restricted core set of thirteen specific fonts, that contain 
>well-known and trivial deviations in the construction of those font names, 
>and that most often are encoded in WebCGM's default character encoding of 
>ISO 8859-1.  The rules may be less useful outside of that intended scope 
>of normalization."
>
>QUESTION 1:  I think that is more accurate, yes?
>
>QUESTION 2:  If "yes", then... should 2.1 font-sub stuff contain a 
>parameter to control whether or not that core-13-deviant normalization is 
>applied?
>
>Regards,
>-Lofton.
>
>
>
>
>
>
>
>

Received on Tuesday, 16 December 2008 23:36:28 UTC