Embedding webfont: a real-life example

[Gérard Talbot]

Hello all,

When webfonts are used for purely cosmetic/ornemental reasons and on a
large scale, I don't agree.

When webfonts are used because Unicode support among browsers for a
particular language (like a Canadian Aboriginal language) is inexistent
(see
http://www.alanwood.net/unicode/unif...syllabics.html
),
when fonts for such language are rare/difficult to find, then I agree.

The best solution is still/nevertheless to
- get/download the font file and then install it permanently on the
user's system; webfonts are just a temporary, not-optimal workaround
- have browser manufacturers create language pack for their browsers
- have browser manufacturers create localized versions of their browsers

Do we all agree? Am I making some sense?

Here's a simple demo for Inuktitut language using nunacom font; you'll
need MSIE 5+.

http://www.gtalbot.org/DHTMLSection/...moNunacom.html

Real-life example: http://www.eenadu.net/

Gérard
--
remove blah to email me

[Alan J. Flavell]

On Mon, 3 Oct 2005, Gérard Talbot wrote:

When webfonts are used for purely cosmetic/ornemental reasons and on
a large scale, I don't agree.
But that seems to be the only specification-conforming possibility, as
far as you've managed to show usenet so far.
When webfonts are used because Unicode support among browsers for a
particular language (like a Canadian Aboriginal language) is
inexistent
I've asked you before, when you posted the same sentiments to
c.i.w.a.site-design, to explain what this "support" is supposed to
consist of - beyond getting fonts with the appropriate glyphs, but you
do not seem to have answered. Now you're off on the same non-standard
rant on another group.

Sure, some scripts *do* need additional support: I know this is the
case for Devanagari, as well as, obviously, for Arabic script. I
doubt that merely offering some downloadable font is going to replace
proper typographical support of that kind, though.
(see
http://www.alanwood.net/unicode/unif...syllabics.html
), when fonts for such language are rare/difficult to find, then I
agree.
Why aren't you helping them to find the fonts? As a mere bystander,
it didn't seem hard to find a couple of *real* fonts, though I kept
stumbling on bogus pseudo-Latin-1 fonts that were playing like Symbol,
or worse.
The best solution is still/nevertheless to - get/download the font
file and then install it permanently on the user's system;
The best solution, for supported writing systems, is to publish
*genuine* HTML4, and get *genuine* fonts (i.e Unicode-based) to
support it. This has been HTML for over half a decade now - stop
trying to drag things back into the previous millennium, please.
webfonts are just a temporary, not-optimal workaround
What you've shown us so far has been a dreadful legacy. Seems to me
the "best" solution is already working, while you're just digging
yourself into a deeper hole.
Here's a simple demo for Inuktitut language using nunacom font;
you'll need MSIE 5+.

http://www.gtalbot.org/DHTMLSection/...moNunacom.html
I say again: your HTML contains ASCII characters, NOT Canadian
Syllabics. That makes it non-standard, and you have no right to go
bragging about "web standards" on that page.
Real-life example: http://www.eenadu.net/

Advertises Windows-1252 encoding, renders in Quirks mode, 260 HTML
errors -- this is "real life" indeed (but nothing to be bragging
about).

Even worse, the source code contains not merely encoded characters
(which, if advertised as x-user-defined instead of Windows-1252, might
be marginally excusable), but the following bogosities (sample only):

ÆOÕ-³Ä-X¾-.ä©ü,
¹K-¯Ã-¹-X¾Üªý,
ÆNÕ-ÅÃ-¦ü-¦--Í.

and so on. The meaning of which is unambiguous in standard HTML, and
it sure does *NOT* represent what these lot want it to mean.

The consequence of which is, they're building a legacy of non-standard
pages which can't be browsed on specification-conforming browsers, at
least not without a fraudulently-constructed font (which is sure to
cause problems elsewhere).

If you know how to make downloadable fonts work with standards-
conforming HTML4, feel free to teach us. If you can't, then please
desist from this retrograde stuff.

[Gérard Talbot]

Alan J. Flavell a écrit :

On Mon, 3 Oct 2005, Gérard Talbot wrote:

When webfonts are used for purely cosmetic/ornemental reasons and on
a large scale, I don't agree.

But that seems to be the only specification-conforming possibility, as
far as you've managed to show usenet so far.

I'm not sure I understand why you say that.

When webfonts are used because Unicode support among browsers for a
particular language (like a Canadian Aboriginal language) is
inexistent

I've asked you before, when you posted the same sentiments to
c.i.w.a.site-design,

Sorry, I missed your post. Jukka had a followup-to (which I didn't
notice when hitting the post button of my email software) and I didn't
subscribe (and don't want to either) to c.i.w.a.site-design. I'll get
back to this later... really busy now...

to explain what this "support" is supposed to consist of - beyond getting fonts with the appropriate glyphs, but you
do not seem to have answered. Now you're off on the same non-standard
rant on another group.

Sure, some scripts *do* need additional support: I know this is the
case for Devanagari, as well as, obviously, for Arabic script. I
doubt that merely offering some downloadable font is going to replace
proper typographical support of that kind, though.

We agree then. Temporary downloadable webfont trick does not solve the
problem: it's a workaround which is justifiable for countries, regions
where the main language is not well supported by browsers, not even well
supported in language packs, localized releases, etc. Are there
keyboards available in Inuktitut? Probably not.

(see
http://www.alanwood.net/unicode/unif...syllabics.html
),

when fonts for such language are rare/difficult to find, then I
agree.

Why aren't you helping them to find the fonts? As a mere bystander,
it didn't seem hard to find a couple of *real* fonts, though I kept
stumbling on bogus pseudo-Latin-1 fonts that were playing like Symbol,
or worse.

The best solution is still/nevertheless to - get/download the font
file and then install it permanently on the user's system;

The best solution, for supported writing systems, is to publish
*genuine* HTML4, and get *genuine* fonts (i.e Unicode-based) to
support it. This has been HTML for over half a decade now - stop
trying to drag things back into the previous millennium, please.

I don't understand your irritation here.

webfonts are just a temporary, not-optimal workaround

What you've shown us so far has been a dreadful legacy. Seems to me
the "best" solution is already working, while you're just digging
yourself into a deeper hole.

Here's a simple demo for Inuktitut language using nunacom font;
you'll need MSIE 5+.

http://www.gtalbot.org/DHTMLSection/...moNunacom.html

I say again: your HTML contains ASCII characters, NOT Canadian
Syllabics.

My os does not use those syllabics: no language pack for XP, no Unicode
support. That too is a real-life situation.

That makes it non-standard, and you have no right to go bragging about "web standards" on that page.

Real-life example: http://www.eenadu.net/

Advertises Windows-1252 encoding, renders in Quirks mode, 260 HTML
errors -- this is "real life" indeed (but nothing to be bragging
about).

Yes, yes. I know that. But at the time the web author created his site,
the font support for his language was rather rare in his country.
Someone asked for a real-life example: then this is one with real-life
issues, imperfections, etc. I know of a few others like that, with lots
of errors of all kinds.

Even worse, the source code contains not merely encoded characters
(which, if advertised as x-user-defined instead of Windows-1252, might
be marginally excusable), but the following bogosities (sample only):

ÆOÕ-³Ä-X¾-.ä©ü,
¹K-¯Ã-¹-X¾Üªý,
ÆNÕ-ÅÃ-¦ü-¦--Í.

and so on.

Yes, yes, yes. But that was not the point of my post. It's a badly
authored webpage but a real-life authored webpage too trying to deal
with a real-life font problem.

Gérard
--
remove blah to email me

[Alan J. Flavell]

On Mon, 3 Oct 2005, Gérard Talbot wrote:

Alan J. Flavell a écrit :

On Mon, 3 Oct 2005, Gérard Talbot wrote:

When webfonts are used for purely cosmetic/ornemental reasons
and on a large scale, I don't agree.
But that seems to be the only specification-conforming
possibility, as far as you've managed to show usenet so far.

I'm not sure I understand why you say that.

Because your example contained ASCII characters, and so the only
specification-conforming implementations will be those which display
ASCII characters, in some cosmetic/ornamental form or another. It's
quite elementary - but you've so confused yourself with non-standard
stuff that you've lost sight of that, it seems.
We agree then.
hardly...
Temporary downloadable webfont trick does not solve the problem:
I agree to that part! Worse, it re-creates an old problem all over
again - that was exposed back in 1997 if not earlier.
it's a workaround which is justifiable for countries, regions where
the main language is not well supported by browsers,
I can't agree. It's a dead-end which produces useless legacy content,
in the interests of avoiding any better solution.

Neither of the examples you've offered so far appear to be justified
now.
not even well supported in language packs, localized releases, etc.
It doesn't need as much messing around as non-standard legacy stuff,
which, nota bene, is *incapable* of being browsed on a
standard-conforming browser (without giving it fake fonts, which will
surely cause trouble somewhere down the line).
Are there keyboards available in Inuktitut? Probably not.
I don't see how this supports your defence of non-standards.

Why aren't you helping them to find the fonts? As a mere
bystander, it didn't seem hard to find a couple of *real* fonts,
though I kept stumbling on bogus pseudo-Latin-1 fonts that were
playing like Symbol, or worse. The best solution, for supported
writing systems, is to publish *genuine* HTML4, and get *genuine*
fonts (i.e Unicode-based) to support it. This has been HTML for
over half a decade now - stop trying to drag things back into the
previous millennium, please.

I don't understand your irritation here.

No, you really don't, do you?

You're encouraging the continued generation of non-standard source
content, that can't be browsed on a conforming browser with conforming
fonts, that can't be indexed by indexing engines, that can't be
processed by anything other than the non-standard fonts on which it
depends - and you don't see why this is a problem.

I say again: your HTML contains ASCII characters, NOT Canadian
Syllabics.

My os does not use those syllabics:

Your OS is not the issue.
no language pack for XP,
Those who need minority content should not expect to be spoonfed -
especially not with nonstandard rubbish which will cripple them for
dealing with the real thing.
no Unicode support.

I ask you for the third time now, what more do you need in the way of
"support" than to install a suitable font? (e.g code2000, if you
can't or won't afford anything better). You still haven't answered
that, other than to admit (AIUI) that your nonstandard kludge doesn't
provide that kind of support anyway either.

As for the other web page which you cited, you omitted to say which
language it was (I don't really recognise the languages and scripts of
the subcontinent, I have to admit). If it was Telugu, then Google
think it's already working: http://www.google.com/intl/te/ , and
certainly I get a convincing display on Mozilla when I access it.
utf-8, of course. I'd say google knows better than you (even if I'd
have to deplore their apparent unwillingness to write syntactically
valid HTML).

Please, stop promoting this obsolete non-standard legacy stuff from
the last millennium, which does nothing better than to disempower
those who fall under its spell.