W3C home > Mailing lists > Public > www-style@w3.org > August 2003

[CSS21]-relevant: Opera-bug 122147 (font-sizes)

From: Jörg Hartmann <jhartmann@aquilacoop.de>
Date: Sun, 3 Aug 2003 15:30:14 +0200
To: <www-style@w3.org>
Message-ID: <009e01c359c3$5fe7afe0$b97ba8c0@PAQUNB01>

Strange. My next message
(http://lists.w3.org/Archives/Public/www-style/2003Aug/0010.html) _did_ get
onto the list, but this one didn't. Well, maybe this instance does ...

-----Original Message-----
From: Jörg Hartmann [mailto:jhartmann@aquilacoop.de] 
Sent: Sunday, August 03, 2003 2:59 PM
To: 'fahrner@pobox.com'
Cc: 'Håkon Wium Lie'; 'www-style@w3.org'
Subject: RE: [CSS21]-relevant: Opera-bug 122147 (font-sizes)


Hi Todd, first of all thanks to you for caring about reasonable font-sizes
and readability and putting so much time and effort into this subject, for a
couple of years now, if I notice rightly. I would have loved to discuss your
thoughts/proposal or the topic in general privately first, and I'm sure we
would agree on basic logic and not waste our time in something which can be
interpreted as a public dispute. But you know: the proposal is in the CSS21
Working Draft now, and time is running shortly, so we have to deal with
publicity somewhat while trying to stay in a conversation that focuses on
the subject (and its inner logic) itself. My apologies for being somewhat
"nervous" in view of a possible W3C-REC which would harden confusion
(different steps within the interval, too small ones --and too many-- below
normal, and at the same time too big ones --and too few-- above normal)
instead of erasing (or at least easing) it. Which is surely not your guilt,
to the contrary, if anything is "guilty" then it's the fact that you were
rather alone most of the time with your commitment. So once more: Thanks
Todd.

> I happen to agree with you that it is hardly useful to have 3 steps
> below the most common or body size in documents.

I know you would. This is why actually you propose _simply not to use_ (in
the h1-h6 hierarchy) one of these steps (x-small). But why have it then? Or
more precise: why not kill it simply?

> But then why not
> simply, as either an author or user (in user stylesheet) set the root 
> size to "small" or even x-small, and go from there?

As a user? How do I set classes to my own preferences if the author did set
them to medium, small, ...? Well, if he did only (correctly) use h1, p and
the like, then it could work. But if he used absolute values (which your
proposal is about) for differentiating his structure, then the user is
simply not able to do anything beyond multiplying _all_ font sizes with the
same percentage. Isn't he?

"Simply set the root size to small or even ..." is really not what you
propose with your font-size interval. If there are (and shall live on)
absolute values for the font-size property, then they should be reasonable.
Why do you propose these values (to let live on - in a confusing system)
when at the same time you advide authors and users to simply not use them?
Why don't you propose then to simply eliminate them altogether - within the
spec (CSS21)? If you did propose that --to simply _eliminate_ absolute font
size values (or to create a completely new, really reasonable system,
without confusion, which indeed would allow users to set the smallest
font-size's pixel-size in their user stylesheet and generate all other sizes
on that base)-- I would agree. But that's not your present proposal. That's
not what I read in Chapter 15.7 of the January Working Draft.

> Jörg, as I opined in a private reply to your concerns
> earlier, I think 
> the problems you describe would best be addressed in browser 
> prefs UI,

Well, but this is opposite to your proposal. You do _not_ propose that the
browser's UI should enable the user to set reasonable steps and intervals of
visual font-size differences, to the contrary you propose a new (strict)
spec for those steps and intervals which hardens confusion by letting there
be "different steps within the interval, too small ones --and too many--
..." (read above).

> not by trying to make the initial value of font-size (medium)
> something 
> other than the legacy HTML font size value of "3"; i.e., the 
> value that 
> obtains in the absence of any contrary direction.

Well, if _that_ is a proplem for you, then why not simply change initial to
small (within the spec ->CSS21)? And to oppose possible confusion (until now
the stylesheets do not link to a DTD of the relevant CSS-version) just use
new, other words for the absolute sizes (why, for instance, not just 1 to 7
as in old html?).

You probably will agree that nothing really stands for "medium=normal" -
except that "medium=initial" (within CSS1 and CSS2, _not_ within natural
law!) could falsely imply that. "Medium" is "in the middle". And as you
wrote yourself, there's no rational for "in the middle between 3 smaller and
3 bigger sizes" would be the right place for "normal volume text".

> Users should no
> longer be asked to specify a single font size preference, but instead 
> should specify the very smallest size they care ever to read at all.
> That size should map to xx-small, with the other sizes above it being 
> generated if not set directly.

Well, that's not really such a big buseness as long as the interval between
sizes is defined. It's just a question for UI recommendations to advide
size="1" or xx-small (or whatever is the smallest size) to be what users can
directly specify within the UI (while the other sizes are calculated based
on the spec's interval). If the spec tells us _anything_ about the
percentage between the available sizes (smallest, "normal", "medium",
"initial", whatever), which does _every_ spec for absolute sizes (so CSS1
did, CSS2 did, and you propose a new variant for CSS21) then it's basically
irrelevant which of these sizes is the one the user directly alters within
the UI (since all are always calculated based on a given interval).

This means also: Users do _always_ specify "a single font size preference",
be it "normal" ("usual body text size", not necessarily the same as
"initial", while at the same time "initial" --as well as the whole system
and its terms-- can be altered from one spec version to another - you
remember) or "smallest" or whatever.

In reality, many browsers allow the users to do both: to set the "normal"
text size value _and_ to set the "smallest ever" size.

> Perhaps show a live preview of the
> entire keyword range, too, to communicate the reality that this is a 
> system, not an isolated value.

You mean I should create a sample sheet and document? Well, at least I
describes some in detail. Maybe you're right and I should also create them
by myself and put the links up here, just for those busy people (on the
list) ...

Todd, you see: you _got_ your proposal into the Working Draft. Which means
it's possible to change things. And believe, it's also possible to really
change (and cure) the whole system, not only to patchwork around ist most
visible wrongdoings. Instead of a workaround for the misunderstandable,
misimplemented CSS1/CSS2 absolute font-size property values, we should ask
for a real solution. And if this means that medium is no longer initial (or
maybe there's no medium anymore _at all_), then be it. If you got a
workaround into the spec's working draft, you'll get a real solution into
it, too. Don't you think? ` Yours for now, Jörg
Received on Sunday, 3 August 2003 09:29:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:23 GMT