Re: [css3-text] Appendix F not needed

Koji Ishii wrote:

> > But if the spec refers to Unicode it's not needed. I think it
> > makes much, much more sense to refer to a Unicode spec for
> > character properties, both from a maintenance perspective and from
> > a domain knowledge perspective.  The CSS group should defer to
> > Unicode on these matters unless there is a very strong reason not
> > to.  Given that UAX 14 has explicitly referred to the line-break
> > property, it makes much more sense to refer to the 'CJ' class
> > directly.  If we want to quibble about the exact set of characters
> > included, we should raise an issue with the Unicode folks and not
> > create our own ad-hoc character classes unless it's absolutely
> > necessary.
> 
> I disagree. Unicode usually defines the baseline for plain text or
> when no other information is available, and usually recommends
> applications or higher-level protocols to tailor as appropriate.
> Since CSS is one of the most common higher-level protocols, CSS
> should tailor Unicode definitions as needed.
> 
> In this case, there are several code points that change their
> behavior by the information Unicode cannot have such as script, and
> there are demands for more features such as ''loose'' than Unicode
> would like to support today, I think tailoring makes more sense than
> to ask Unicode to do all the work we need.

Koji, I'm not saying line breaking behavior should be defined only in
UAX14 or that you shouldn't include exceptions for specific codepoints
in a property definition, I'm simply saying the need to list a set of
codepoints is unnecessary since they are already defined in UAX14.

Regards,

John Daggett

Received on Monday, 10 December 2012 01:39:14 UTC