Version reviewed: http://www.w3.org/TR/2008/WD-webcgm21-20080917/WebCGM21-Config.html#ACI-fontmap
Lead reviewer and date of initial review: Richard Ishida, Nov 2008
Subject lead in: [WebCGM2.1]
This review is focused on section 9.3.2 specifically. These comments have not yet been endorsed by the I18N Core WG. The "Owner" column indicates who has been assigned the responsibility of tracking discussions on a given comment.
We recommend that responses to the comments in this table use a separate email for each point. This makes it far easier to track threads. Click on the icons in the right-most column to see email discussions.
substitutionList="CDATA" should be bolded
"The list syntax and normalization are derived from the specifications of CSS 2.0 [CSS20]. In particular, the values and syntax of the substitutionList attribute are derived from CSS's definition of the font-family property "
It is unclear where these rules are in CSS2, ie. whether there are more than in the font-family section, given the use of the phrase 'in particular'. If so, we couldn't find where.
Why is the normalization for cgmFont different from that for substitutionList?
|4||18.104.22.168||Stripping out all whitespace||
For cgmFont normalization, do you mean "stripping out all whitespace" or normalizing white-space to a single space, as per substitutionList?
"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?
Normalization for string comparison should include conversion to a Unicode normalization form, to eliminate issues related to precomposed vs. decomposed characters and issues related to ordering of multiple combining characters.