From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Tue Mar 30 2004 - 20:19:34 EST
From: "Kenneth Whistler" <kenw@sybase.com>
> > Users can define only those properties which the
> > software that they are using allows them to define. Your argument here
> > completely ignores the distinction between users and software
> > developers.
>
> No it doesn't. I am well aware of the distinctions between end
> users, application developers, OS platform developers, and
> basic library implementers. I have, at one point or another,
> been in all of those shoes.
Would adding those alternate properties in PUA fonts violate Unicode? I don't
think so because it is application-specific, but this does not mean that no
standard can be developed here to carry those information that Unicode refuses
to integrate in its standard.
In fact, your argument to maintain a strong default LTR directionality for PUAs
means that people have no other choice than developing psecific applications
that can bypass this interoperability problem (also because any application that
would use a distinct default for PUAs would, in my opinion, violate Unicode).
So why not adding character properties tables in fonts, and get something that
would please users of those PUAs as they could now start interchanging
documents, not by using only Unicode encoded plain-text with PUAs, but by
agreeing simply on the best font to use to render or parse a document using
those PUAs. OpenType tables in fonts are extensible, and if such standard table
formats were specified, they could finally be recognized in various renderers to
get what users want. Or used in text parsers if instructed to get the properties
tables from a font (even if this is not for rendering purpose).
An alternative could be to develop a compact interchange format for properties,
and to make this file format optionally embeddable directly in fonts (so that it
becomes easier to perform text handling without rendering).
Saying that something is private and requires people agreement does not forbids
the developement of a new standard. After all, all standards are born first by
private mutual agreement, and the adoption of the same agreement by more
interested people.
Other interchangeable Unicode properties formats are also needed, notably for
the localized collation tables. Will Unicode work to prohibit such developments?
It would be unfair for those that want to use the freedom given to them with
PUA. Private use does not mean that this can't be interchanged. It's just that
it has no standard behavior defined by Unicode itself, and Unicode makes this
statement of allowing PUAs as a guarantee offered to users to allow such
separate standard development without conflicting with other Unicode works.
This archive was generated by hypermail 2.1.5 : Tue Mar 30 2004 - 21:07:00 EST