RE: EOT Lite (Compatibility Web Type) - sample web pages

Tuesday, September 08, 2009 Jonathan Kew <jonathan@jfkew.plus.com>:

>Sorry, Mozilla "try-server" builds expire after two weeks, so this is no
longer online.

Jonathan,

Color me confused and concerned. I've been using the Minefield build with
EOTL support all weekend, freshly downloaded from here:
https://build.mozilla.org/tryserver-builds/jkew@mozilla.com-try-cdcb63f9c65c
/try-cdcb63f9c65c-win32.zip
But I had a bit of trouble with using the raw files all by themselves and,
first, had to use the windows installer package of Minefield located here:
https://build.mozilla.org/tryserver-builds/jkew@mozilla.com-try-cdcb63f9c65c
/install/sea/try-cdcb63f9c65c-win32.installer.exe
I installed it under a separate profile to prevent messing up my other FF
installs, overwrote the contents of c:\program files\Minefield with the
zipped files, and it's been working great.
Thanks. What a pleasure not to have to bounce back and forth between IE and
FF.

I've just written about EOTL (or should we begin calling it CWT at this
point?) at http://readableweb.com and was going to follow up in a few days
with instructions for those looking to install Minefield and if those files
are going south I'd really like to know. If they do disappear, are the ones
I've downloaded redistributable?
Further along these same lines, I understand from
https://bugzilla.mozilla.org/show_bug.cgi?id=507970
that you've done a build with WOFF support. Great. However, although I know
you've published the spec, is there, as yet, a tool for converting TTF or
OTF files into WOFF files for testing?

And lastly in this vein - the $64,000 question - how soon can we expect a
build with support for WOFF and EOTL(CWT) side-by-side?

>I took a quick look at the "Mayberry" page (sample 3), and I notice  
>that the "preview" code displayed on the page does not in fact  
>correspond to the CSS that is actually used for the extended family.  
>For the latter 4 faces, the CSS actually uses separate family names,  
>rather than using the true family name and relying on the descriptor  
>values to identify the proper faces. Is this an oversight, or was it a  
>hack that turned out to be necessary in order to get the desired  
>rendering?

Quick catch. I'm just getting into the CSS syntactical issues and IE's
behavior in regards to EOTL as opposed to the standards-compliant syntax and
behavior. However, my first impression of the Mayberry Sample 3 page is that
those involved decided to present the CSS for each weight as if that weight
was the only Mayberry font being used on the page.
In that, I think that Ascender's page might be somewhat misleading to those
who would use it as a guide to working with multiple weights of the same
name using EOT in IE, let alone squaring that with what Firefox and other
standards compliant implementations can and will do with the same syntax.
But frankly, font-linking as a practical proposition is so new, I don't know
of anybody who's mastered all the ins and outs.

I'm also under the impression that IE has some issues with numerical weight
properties, period, which might complicate matters. Time to go digging and
get the facts.
Of course, this "differential" between FF's implementation and IE's
@font-face implementation and/or other CSS shortcomings needs to be
thoroughly understood and explained.

As I said before, this issue seems to be getting solved by the font services
folks, so I assume - hell, *I know* - that DIY folks like me will also
figure it out and codify what's best practice. Something akin to
"@Font-Face: A Guide To The Perplexed" can't be far off.
And insofar as font-linking using EOT files in IE goes back to when most web
developers I deal with were still in puberty, the word "hack", I feel, is
unfairly negative given the long history. We've known since day one that
workarounds were going to be necessary. But the benefits are great. We've
had to work around worse. And I can't believe IE won't get @font-face in
line with the spec in IENext. (Sylvain?)

Please, please let us know about the status of WOFF tools, WOFF and
EOTL(CWT) side-by-side, etc. as soon as is practical. I don't want to do
anybody a disservice by publishing incorrect info.
I've heard of the five second rule
(http://en.wikipedia.org/wiki/Five-second_rule), but the two week rule is a
new one! ;)

Regards,

rich




-----Original Message-----
From: www-font-request@w3.org [mailto:www-font-request@w3.org] On Behalf Of
Jonathan Kew
Sent: Tuesday, September 08, 2009 12:56 PM
To: info@ascenderfonts.com
Cc: www-font@w3.org
Subject: Re: EOT Lite (Compatibility Web Type) - sample web pages

On 7 Sep 2009, at 23:45, Bill Davis wrote:

> We made a variety of web pages to demonstrate EOT Lite fonts in  
> action. They showcase simple headline fonts, extended font families  
> and fonts with larger character sets.
>
> http://www.ascendercorp.com/web/web-font-examples/
>
> The pages can be viewed in Internet Explorer and the Firefox  
> minefield version that Jonathan Kew posted a few weeks ago (but the  
> download link does not appear to be active).


>
> I know some folks have noted issues/challenges with formatting font  
> families in IE - not sure if our approach is 'correct' but it  
> appears to work fine in our testing. We made it easy to preview the  
> code. Look forward to comments from you all.
>

I took a quick look at the "Mayberry" page (sample 3), and I notice  
that the "preview" code displayed on the page does not in fact  
correspond to the CSS that is actually used for the extended family.  
For the latter 4 faces, the CSS actually uses separate family names,  
rather than using the true family name and relying on the descriptor  
values to identify the proper faces. Is this an oversight, or was it a  
hack that turned out to be necessary in order to get the desired  
rendering?

JK

Received on Tuesday, 8 September 2009 20:21:30 UTC