Re: relation between unicode and font

From: Jukka Korpela (
Date: Sat Jan 06 2001 - 02:13:08 EST

On Fri, 5 Jan 2001 11:59:27 -0800 (GMT-0800), you wrote:

>With that said, this type of "bug" is actually by design for many languages
>that have speakers who are not using Unicode. When you look (for example) at
>Indic languages on the web, the vast majority of them are handled by this
>type font hack.

It is true that various font hacks are widely used. This just emphasizes
the problem and the need for more reasonable methods, especially since
the hacks do not work in the worldwide environment. What I tried to make
clear that the Unicode approach is completely different and specifically
abhors the idea of changing the _meaning_ of data by font selection.

>So I am hesitant to call it a BUG. Its just the encoding that people will

It's still a browser bug if the document is declared or implied to be
ISO 8859-1 encoded and <font face="Tikkana">A B </font> causes something
else than A B to be displayed.

It would be technically correct in _some_ sense to have a document sent
over the network with information (in HTTP headers, E-mail headers, or
elsewhere) that declares the encoding as x-something, i.e. a "private"
agreement, and the recipient is expected to tune his program to act
accordingly. Then, in principle, the <font> tags would be just what they
are supposed to mean, and ultimately mean: a suggestion to use some
particular glyphs for the presentation of some characters, the
_identity_ of which is established elsewhere.

But that would not make much difference in practice. It is much more
realistic to expect that Unicode will be supported to some reasonable
extent in the near future than to register various 8-bit encodings
(after all, by Internet protocols, the encodings, or "charsets", should
really be registered at IANA) and to expect that they will be actually
supported by various programs worldwide. It is rare to see even all the
ISO 8859 encodings (or even those of them that have been defined as
international standards years ago) adequately supported by Web browsers
and software, even in new versions.

Qui nescit tacere nescit et loqui

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:17 EDT