From: jameskass@att.net
Date: Tue Feb 10 2004 - 23:02:16 EST
Chris Harvey wrote,
>... I want the
> examples on my site to be legible (dot accents non-spaced in the middle of
> syllabics instead of above them aren't really acceptable), and I want the
> characters to look like what speakers are familiar with, otherwise they may
> very well choose not to use the font, keyboards, etc.
>
> My aim is that people can type their own language on the computer they have
> now. Once OpenType is available on my machine and others, I will release
> fonts which have OpenType tables, calling the same glyphs that are now in
> the PUA. This way, I am trying to make some humble attempt at backward
> compatibility. But for now, if people cannot use the OpenType
> substitutions, what else should I do?
>
> I am building specific fonts for specific languages, but I wanted one font
> that would display the lot. That way, if someone wanted to use
> languagegeek.com, they would only have to download one font, instead of one
> per language.
These are all laudable goals with understandable intentions. As far as
*characters* which aren't yet encoded, the PUA really seems to be the
only method.
Since you asked, however, an alternative to the current approach would
be to:
* Encode the pages as compliantly as possible.
* Offer the one font to fit all the pages while awaiting either
language-specific fonts or OpenType technology availability.
* Note on the pages that the one font aims to cover all syllabics, but that
language-specific variants exist which can't yet be covered in a single
font due to technological limitations.
* Use any combining dots and so forth from the COMBINING
DIACRITIC range. (A font like Code2000 won't display these
combiners well due to technology limitations, but, so what?
In *your* font, you can place the combining glyphs so that
their default position is acceptable and won't overstrike the
base glyphs.)
An advantage to doing something like the above is that backwardness
isn't being perpetuated under the guise of backwards-compatibility.
Another merit is that text (aside from necessary PUA matter) is
correct, compliant, interchangeable, and permanent. Parsers,
search engines, indexing operations, and all the rest, will work
as they should.
A disadvantage of the current approach is that users may be too
easily tempted to also generate text, data, and web pages using
a proprietary encoding. In the long run, many might view this
a something other than a favor to the user communities.
Best regards,
James Kass
This archive was generated by hypermail 2.1.5 : Tue Feb 10 2004 - 23:48:43 EST