W3C home > Mailing lists > Public > www-style@w3.org > March 2010

RE: [css3-fonts] font-specific feature handling

From: Richard Fink <rfink@readableweb.com>
Date: Sat, 20 Mar 2010 21:14:56 -0400
To: "'Tab Atkins Jr.'" <jackalmage@gmail.com>
Cc: "'Christopher Slye'" <cslye@adobe.com>, "'www-style'" <www-style@w3.org>, "'Jonathan Kew'" <jonathan@jfkew.plus.com>
Message-ID: <003401cac893$ebcc4160$c364c420$@com>
Saturday, March 20, 2010 2:09 PM <jackalmage@gmail.com>:

Tab,

I can tell when I'm licked! So ends a thought experiment. Perhaps script is the best vehicle for fallback in such cases.
Or, as Jonathan Kew wrote:

>If the precise visual appearance of the text is truly critical to the message,
>then I would suggest that the author should in fact be using some kind of graphic -- what
>is primarily in focus is the graphic appearance, more than the textual content.

To get back to the initial spark concerning font-variant and contextual alternate glyphs:

Jonathan Kew wrote:

>>And if font designers or web authors are using font features such as contextual alternates to access glyphs that
>>fundamentally alter the meaning of the text being read, then.... well.... they are abusing the technology, and
>>thoroughly deserve any resulting confusion. That's something we can frown upon, but we cannot legislate it into
>>impossibility -- any more than we can prevent people using @font-face to deliver fonts where the character-to-
>glyph mapping has been scrambled so as to completely subvert the normal reading of the text.

Call it "abusing the technology", call it sabotage, call it, as you did previously in reference to images without alt tags, bad authoring - there is a limit to which this can and should be legislated by the standard. Where is that line? Is it the needs of the many outweighing the needs of the few? I don't think there's a bright line.
But what I do know is that in the case of contextual alternates the author can choose one of two forks in the road: 1) They accept more verbiage with no risk of the alternates falling through to the fallback fonts.
2) There is less verbiage but there is risk of fall-through to the fallback fonts.
To me, it makes little sense to penalize the author who chooses #2 sensibly and responsibly because someone else might not be sensible and responsible - especially since #1 - the first, no-risk road is being made available to them.

Agreed?

To cap off my "no fallback" - some clarifications:

>Why do you want to force upon the visitor a blank page?
I don't, really. But what alternative there would be, I'm not sure. Look, we *want* to encourage the use of fonts instead of text-as-images, right? Do we agree it's the more accessible, more elegant option? Well, in some situations, say a logo for instance, a bitmap or vector image with the HTML text as alt text constructed on the fly using Canvas or VML as the Cufon javascript text replacement library does, might make sense.
I am concerned about there being a jolt in the way a page displays between UA's that support @font-face and those that don't. But once again, script might be the best answer because at least the fallback font can be adjusted to a specific font-size - or even font choice - that's closer to the webfont that may be unavailable. That part of the fallback stack - the inability to tweak each font in the stack - has always bugged me, and I'm not the only one.

>In the future, this can be accomplished purely through
>stylesheets in an accessible manner with the content property.  Put
>text in your HTML, substitute an image with the content property, and
>not only do you maintain perfect accessibility, but it automatically
>degrades to something usable if a sighted user is having network
>problems. 

Agreed. Problem solved.

>It is so. There is such a thing as too much freedom. Your freedom to
>swing your fist ends at my face, etc.

Can't say I’m in love with this imagery. Can we settle for shouting "Fire!" in a crowded theater instead? :-)

Regards,

Rich

-----Original Message-----
From: Tab Atkins Jr. [mailto:jackalmage@gmail.com] 
Sent: Saturday, March 20, 2010 2:09 PM
To: rfink@readableweb.com
Cc: Christopher Slye; www-style
Subject: Re: [css3-fonts] font-specific feature handling

On Sat, Mar 20, 2010 at 8:34 AM, Richard Fink <rfink@readableweb.com> wrote:
> I'm surprised by your reaction. It seems like you're objecting to authors having too much freedom.
> Say it ain't so, bro'!

It is so.  There is such a thing as too much freedom.  Your freedom to
swing your fist ends at my face, etc.

> I see fonts as something more fundamental than "styling". And judging by the amount of Flash you see on the web sites of Fortune 500 companies, they do, too. Ask a company that's commissioned a custom font to maintain a consistent brand identity if they would rather the UA display their stuff in Times New Roman or just skip it. I'm betting it will be skip it.

Occasionally there are decision where what is in the business's best
interests is *not* the same as what is in the public's best interest.
I believe that part of our mission as an open standards organization
is to make technology better for *everyone* in general, which means
that in this sort of case, where something is a *pure negative* for
the public and a *marginal positive* for the business, that we should
choose not to pursue it.

> Look, if I go to the trouble of setting up @font-face declarations, ornery cuss that I am, if those fonts don't load, maybe I would prefer failure. What's wrong with that? Why do I have to have the UA/OS force upon me a font I don't want displayed?

Why do you want to force upon the visitor a blank page?  What's
important to you is not the same as what's important to the viewer.
The viewer may appreciate your attention to detail and fine aesthetic
judgement in selecting a font that complements your overall design.
However, that is not why they visited your site.  Unless your site is
an art gallery showcasing fonts, the visitor is there to read some
content you have put there.  Good fonts may make this experience more
pleasant, but fallback fonts won't do anything worse than make it
somewhat ugly.  They can still get what they came for, and that's
what's important to them.

Look at it from this perspective: I could take your argument nearly
word-for-word and use it to justify the 1990s practice of serving up a
website to one browser, and just showing a "Please use XXX" message to
everyone else.  Every justification you have so far given applies
equally to that abhorrent practice, except that the arguments would
have been even *stronger* back then, as they were about much more than
a simple font.  Unless you have something that clearly distinguishes
your desire to show a specific font from those early web designer's
desire to show a specific design, your wish is equally deserving of
rejection.

> You might consider the following as edge cases but they're still food for thought:
>
> How about pages where examining the webfont is the whole point, like the in-browser specimen pages on a site like Font Spring?

Those are best done as images, though supplementing the image with a
live browser-generated specimen is also cool.

> How about pages that are re-creations of old books and documents - along with re-creations of the original fonts?

The font, while interesting to view as it provides a sense of history,
is in no way a necessary part of the document.  A fallback font would
be perfectly serviceable, and often more readable than those early
fonts.

> And when an image doesn't load, which might very well be a background text-as-image, what does the UA fall back to then? Nothing and nobody seems to be losing sleep over it.

Generally, text-as-image in a CSS background is a bad idea.  I
banished that technique from my repertoire when I realized that it was
so much easier to just put an actual <img> in the document with @alt
text.  All the benefits of pretty text, with much easier accessibility
(no more hiding visible text with a negative text-indent).  This does
have the slight downside of me providing "style" through HTML in me
replacing perfectly serviceable text with a prettier <img>, but that's
minor.  In the future, this can be accomplished purely through
stylesheets in an accessible manner with the content property.  Put
text in your HTML, substitute an image with the content property, and
not only do you maintain perfect accessibility, but it automatically
degrades to something usable if a sighted user is having network
problems.

> What specifically is your fear about what would happen if, within @font-face, fallback could be disabled?

I fear precisely that - that fallback would be disabled.

~TJ
Received on Sunday, 21 March 2010 01:15:28 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:25 GMT