Re: Embedding fonts

On Sun, 2 Oct 2005, Gérard Talbot wrote:

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

Something wrong here. This document contains what the HTML
specification defines to be ASCII characters, but they have been faked
to display in a misbegotten downloaded font - which might give a
visual impression of being what the author intended, but fail to
conform with web standards.

Sending it out with iso-8859-10 encoding was amusing, but doesn't
really address the issue at all, since these characters are still
ASCII.

If you're going to create fonts, then for Heaven's sake create them to
published specifications, to support the correct HTML (i.e Unicode)
characters (U+14xx onwards), instead of going and breaking HTML for
your purposes (perhaps we should call that "unocide", SCNR).

It's particularly infuriating that you display "valid HTML" and "web
standards project" icons on the page - "valid" HTML (in the technical
sense) it might be, but the content is bogus in terms of the HTML
specification and Unicode standard.

Unicode support in browsers for Aboriginal Canadian languages is
almost inexistent:

They won't get any better if those who need them, fail to use them!

http://www.alanwood.net/unicode/unified_canadian_aboriginal_syllabics.html

Can you say what this "support" would consist of? Every single
character on that page is displaying a nice-looking glyph in my
Mozilla, in conformance with HTML specifications, and resembling the
glyphs shown at http://www.unicode.org/charts/PDF/U1400.pdf etc.

I see the same at my own test pages from U+1400 to U+1676 inclusive,
http://ppewww.ph.gla.ac.uk/~flavell/unicode/unidata14.html

Screen shots available if you wish...

Oh, did I mention that Alan Wood's page is also displaying just fine
(as are my own test pages) in MSIE? In *that* case, I'm suspecting
that it was adding Japanese support which made the difference, since
lots of interesting non-Japanese characters started working in IE
after I did that. (This is Win/2000 Pro, for the record).

So I say it's already working, whereas your IE-only "solution" for
this non-problem is utterly broken, in HTML terms. You've got a
nerve, really, invoking "web standards project" on such a flagrant
page.

sorry

In our last episode, Gérard Talbot <newsblahgroup (AT) gtalbot (DOT) org>
pronounced to comp.infosystems.www.authoring.site-design:

But it is valid HTML 4.01.

Which is why Alan made the distinction between being valid (in the
technical sense), and being standards-compliant.

Hold it right there. *_You_* are saying "it's already working". You. No
one else is saying that.

Well, because you asked so nicely I downloaded the appropriate fonts,
and both pages in question look fine here (tested in Firefox 1.5beta,
IE6).

http://www.alanwood.net/unicode/unified_canadian_aboriginal_syllabics.html
http://clarkecomputers.com.au/usenet/alanwood.gif [5kB]

http://ppewww.ph.gla.ac.uk/~flavell/unicode/unidata14.html
http://clarkecomputers.com.au/usenet/alanflavell.gif [11kB]

It's already working. :-)

--
Mark Parnell
http://clarkecomputers.com.au

In our last episode, Gérard Talbot <newsblahgroup (AT) gtalbot (DOT) org>
pronounced to comp.infosystems.www.authoring.site-design:

Can you upload a screen shot of this page for me?

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

http://clarkecomputers.com.au/usenet/gtalbot.gif [7kB]

--
Mark Parnell
http://clarkecomputers.com.au

Mark Parnell a écrit :

In our last episode, Gérard Talbot <newsblahgroup (AT) gtalbot (DOT) org
pronounced to comp.infosystems.www.authoring.site-design:


But it is valid HTML 4.01.


Which is why Alan made the distinction between being valid (in the
technical sense), and being standards-compliant.

It is valid HTML 4.01 and CSS 2 compliant as well. I had problems with
doing the webpage and making the text for reasons no one so far
understood. I had to uninstall the font in order to actually test the
page...which is normal procedure, testing. If I have the font installed
on my system, then a webpage with an embedded webfont is not useful.

Hold it right there. *_You_* are saying "it's already working". You. No
one else is saying that.


Well, because you asked so nicely I downloaded the appropriate fonts,

When you download and install the font, everything is ok, everything is
correct. But my original perspective was regarding people who do not
have the font installed..

and both pages in question look fine here (tested in Firefox 1.5beta,
IE6).


http://www.alanwood.net/unicode/unified_canadian_aboriginal_syllabics.html

http://clarkecomputers.com.au/usenet/alanwood.gif [5kB]


http://ppewww.ph.gla.ac.uk/~flavell/unicode/unidata14.html

http://clarkecomputers.com.au/usenet/alanflavell.gif [11kB]

Thanks for the screenshots. I saved them. I appreciate your cooperation
here.

It's already working. :-)

Ok, then: uninstall the font and you won't see the glyphs at
http://www.alanwood.net/unicode/unified_canadian_aboriginal_syllabics.html

I'll do another page later... I need to test something that infuriated
and upset Alan Flavell said...

Gérard
--
remove blah to email me

On Tue, 4 Oct 2005, Gérard Talbot wrote:

Let's go with the screen shots then. I want to see them.

For your entertainment, this extract was taken on a machine which
happens to have Code2000 available as proportional font, but does not
have a monospaced font supporting the repertoire:

http://ppewww.ph.gla.ac.uk/~flavell/tests/14.gif

Don't jump up and down yet, though, because (as I report on my page on
the general topic), although MSIE refuses to select a proportional
font as its default, it seems quite willing to honour CSS requests to
use a proportional font in a monospaced context.

My Unicode test pages are deliberately composed *without* CSS
font-face selections, as a diagnostic technique, but in real-life
(remember that? ;-), even MSIE can be persuaded to do something
useful.
http://ppewww.ph.gla.ac.uk/~flavell/charset/browsers-fonts.html

Alan J. Flavell a écrit :

In *that* case, I'm suspecting that it was adding Japanese support
which made the difference, since lots of interesting non-Japanese
characters started working in IE after I did that.

Apologies, my suspicion was wrong. In fact, the characters were
coming from the Code2000 font.

There are also other general fonts which cover this part of the
repertoire, as well as (I gather) specialised fonts which have been
properly built for Unicode (not to be confused with fake fonts that
pretend to be Latin-1, which are evidently around).

Everson Mono (monospaced, shareware) is a general font which covers
this repertoire, to take one example.

On Tue, 4 Oct 2005, Gérard Talbot wrote:

Can you upload a screen shot of this page for me?

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

You mean like this?
http://ppewww.ph.gla.ac.uk/~flavell/tests/NunacomShot.gif

In our last episode, Gérard Talbot <newsblahgroup (AT) gtalbot (DOT) org>
pronounced to comp.infosystems.www.authoring.site-design:

It is valid HTML 4.01 and CSS 2 compliant as well.

There is a difference between being technically valid and complying with
the standards. The former is a part of the latter, but they are not the
same.

When you download and install the font, everything is ok, everything is
correct. But my original perspective was regarding people who do not
have the font installed..

Surely anyone who can actually read it will already have the correct
fonts installed? Sure, I didn't have them installed, but then I can't
read it even with the correct fonts.

--
Mark Parnell
http://clarkecomputers.com.au

On Tue, 4 Oct 2005, Gérard Talbot wrote:

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

You know: *if* you could base your embedded font on *this* version,
which contains *correctly* coded Canadian Syllabics characters, then
you'd be all set. The resulting page would display on any
www-conforming browser that already had access to a suitable font;
while MSIE should be able to use the embedded font, even if it didn't
have access to a suitable font locally.

I thought I'd try this myself, so I downloaded the MS WEFT3 tool,
which I've never used before, and tried it against various fonts.
Several which covered the needed repertoire were prohibited for
embedding, but the "Aboriginal Sans" font seemed to be eligible.

Initially, the tool seemed to have no idea which font to use, but then
I spotted the comment that the WEFT page analysis involves MSIE
somehow; so I went into MSIE's configuration menu and configured it to
default to using "Aboriginal Sans" for Canadian Syllabics - after
that, the WEFT3 tool analysis phase seemed to have worked out
correctly that it could use this font for its embedding purposes.

Unfortunately, after this point things seem to go amiss. There seems
to be something wrong with the "Fonts to embed" dialog (can you
compare the screenshot[1] with what you're used to, please?), and when
the software announces "Creating EOT files.", I have to report that no
such files seem to be created - neither where I expected them, nor
anywhere else that a search can find.

So at this point I'm stuck. Either this is due to my unfamiliarity
with the software, or it really can't cope with what I'm trying to
persuade it to do. Unfortunately, their demonstrations don't seem to
cover this kind of usage - in their examples (as far as I could tell)
MS seem to be using only a pedestrian character repertoire, and
concentrating on getting different cosmetic results.

comments, please?

[1] The WEFT3 user interface appears to consist of unresizeable
windows which are too small for their contents, and there are texts
which refer to controls that aren't in fact displayed in the windows:
in this example, it says "Use the Embed/Don't Embed button", but there
is no such button to be seen:
http://ppewww.ph.gla.ac.uk/~flavell/tests/fontstoembed.gif

Mark Parnell a écrit :

In our last episode, Gérard Talbot <newsblahgroup (AT) gtalbot (DOT) org
pronounced to comp.infosystems.www.authoring.site-design:


It is valid HTML 4.01 and CSS 2 compliant as well.


There is a difference between being technically valid and complying with
the standards. The former is a part of the latter, but they are not the
same.

I understand the difference. Alan F. was furious because the document
claimed to be valid HTML. And, so far, everyone agrees that the document
was valid HTML and is using valid HTML.

I mishandled the character encoding, usage of character entities instead
of ascii. Acknowledged. The web standards org gif is a gif included in a
template; I had troubles, problems when doing that page... and that
standards.org gif was really the last thing on my mind. Why do I have to
go into lengthly details about this?

When you download and install the font, everything is ok, everything is
correct. But my original perspective was regarding people who do not
have the font installed..


Surely anyone who can actually read it will already have the correct
fonts installed? Sure

.... unless webfonts are used or an image of text is used. That was my
point.
How do you tell in a webpage the users of these language to use (explain
and/or download and/or install and/or use) fonts supporting their
languages, alphabets in webpages? Do you write them that in plain
English in a webpage?

, I didn't have them installed, but then I can't

read it even with the correct fonts.

Microsoft could have made a big difference for all of us. By installing
by default a reliable unicode font by default on any/all windows
installation. That way, that would sovle a lot of problems.

Gérard
--
remove blah to email me

On Wed, 5 Oct 2005, Gérard Talbot wrote:

I understand the difference. Alan F. was furious because the
document claimed to be valid HTML.

Please check the record. I know what "valid HTML" means.

What I was upset by was the presence of the web standards icon (which
was inappropriate), alongside the purely formal "valid HTML" thing -
which, yes, was pedantically correct in and of itself, but distracted
from the *real* violation that was going on there.

And, so far, everyone agrees that the document was valid
HTML

Me too

and is using valid HTML.

Word play.

I'm still interested to see whether this WEFT thingy is capable of
doing the *right* thing, but I haven't managed to persuade it yet.

On a different PC (one that isn't configured with a custom dpi display
setting, which apparently their menu routines can't stomach), I
managed to create what purported to be the .eot file for the real
Can.Syllabic texts, but when I try to use it (with the
syllabic-capable font de-installed), then instead of empty boxes I get
crossed-out boxes.

Incidentally, on wider reading I see that recent versions of Netscape
browser can also do embedded fonts of some kind, even though Mozilla
can't. Presumably it's incompatible with MS ...?