W3C home > Mailing lists > Public > site-comments@w3.org > January 2010

Re: Summarizing the "font-size" issue to make an informed choice

From: Felix Miata <mrmazda@earthlink.net>
Date: Mon, 18 Jan 2010 23:07:05 -0500
Message-ID: <4B552FE9.5060803@earthlink.net>
To: w3c help <site-comments@w3.org>
On 2010/01/12 20:51 (GMT-0600) Ian Jacobs composed:

> > Some initial conclusions based on your comments are:

> >   * I need to fix the CSS to enable resizing in IE. (This might be  
> > done in several ways.)

The simplest "fix" is to permit the users' default size to be used, so that
"re"-sizing is unnecessary.

> >   * I need to review other W3C resources (e.g., QA Tips) where there  
> > is inconsistency with the approach
> >     we take.

In particular, I pointed this out before the recent relaunch:
http://lists.w3.org/Archives/Public/site-comments/2009Sep/0005.html ....

> > I'd like to make an informed choice by understanding the various  
> > considerations. Corrections and additions welcome.
> > Thank you again for your help,

....and again after relaunch:

> > [1] http://lists.w3.org/Archives/Public/site-comments/2009Nov/0048.html

> > Regarding using px to set a base font size

> >     a) It is unnecessary (you can use relative units). However:
> >        - px settings may give you better cross-browser compatiblity

Cross browser compatibility? Ordinary users don't normally open the same page
in multiple browsers looking for "incompatibilities". Web sites do not need
to look identical in every browser. There's no user benefit to this so-called
benefit that web site stylists seem to think important.

> >        - em/percent may give you better resizing capabilities

There's no "may" about it. "Better" includes obviating the need for resizing,
which is ostensibly the most common result of using the users' defaults at
the outset.

Note that some users "resize" in advance, by setting the browser defaults
differently from those issued by the browser vendor. After all, most are
using personal computers, which every public web site stylist should expect
users either to have personalized beyond moving icons around the desktop or
changing desktop wallpaper, or have no need or desire to further personalize.

> >     b) It may pose usability/accessibility issues (when done in a way
> >        that does not allwo resizing).

This is an unfortunate and unnecessary consequence of sizing web page objects
without regard to user settings and preferences, which is exactly what
setting sizes in px and pt do. This disregard is unnecessary, and _rude_.

> >     c) It contradicts W3C's own advice (e.g., a QA tip [1]) not
> >        to mention the advice of other resources.

> >        [1] http://www.w3.org/QA/Tips/font-size

Exactly. Moreover, the W3 site should be a place people looking for authority
on best styling practices can expect to find them implemented.

> > Regarding specifying a base font size other than "the user's default"

> >     a) It is common. The consequence is that if a site doesn't (for good
> >        reasons), the site looks odd.

That's an unfortunate result of trying to be like everyone else without
regard to the sanity of what those others do, and the legacy of significantly
lower DPI computer displays mainstream last century.

> >     b) It is impolite to override preferences explicitly set by users.

Very, and this is also true when these settings are not explicitly set by the
user, because they remain acceptable even though actually set by
applications' authors.

> >     c) Some browsers allow users to set a minimum font size; others  
> > don't.

Note the frequent and undesirable result of minimum size application is that
most or all of an author's contextual text sizing is lost.

> >     d) Users may not know they can adjust font-related settings.

Whether they know they can or not is irrelevant. This statement is frequently
posited as a justification for disregarding user defaults entirely, or
adjusting them downward. It quite _unjustifiably_ assumes there is something
wrong with most user defaults that the stylist can and should do something
about. There is currently no rational basis for making this assumption, or
determining an appropriate amount of "correction".

> > Regarding the user's ability to scale text

> >     a) Some browsers have issues resizing sizes expressed with certain
> >        CSS units (notably px).

Really not just "some" browsers, but only the by far most common one,
Internet Explorer. Page zoom is a way around this for many users in its more
recent versions, but users shouldn't need to resize, which generally only
becomes necessary because the site stylist has done an initial resizing in
the smaller direction.

> > The difficult choice

> > If I understand correctly, we can choose between:

> >   i) Specifying the default font size the way that many other sites do  
> > (which is smaller than the default
> >      used by many browsers).

> >      Pro: consistency across sites

Only a "pro" for those few who compare a page's rendering among more than one

> >      Con: too small for some users, and disrespectful of some who have  
> > set preferences

Disrespectful of some, no. Disrespectful of _all_, yes, because the
preferences many, maybe most, started with, were/are just fine and
needed/need no adjusting.

> > ii) Specifying the default font size to be "the user's default"

> >      Pro: recognizes user should have ultimate control

After all, they're their web page viewing devices, which provide _user_
agents, not author agents, to render the pages.

Attempting to override the user is attempting to thwart a significant
advantage web readers have over print readers, the inherent flexibility that
enables web reading to be more convenient and more comfortable.

> >      Con: will look too large for many users today

Unless the stylist is actually seeing what the users see, he has no yardstick
with which to overall judge as "too" anything, large or small. He doesn't
have my visual acuity, or my viewing distance, or my screen resolution, or my
display size, or my ambient lighting conditions, or any of the other varying
conditions constituting a component of what makes any object a suitable size
for any outside his sphere of direct control.

What the stylist should be doing is sticking to 100% truly relative[1]
sizing, such that all objects on his pages produce desirable relationship to
each other, without any reference to any measurement of unknown and
unknowable size outside his own environment (e.g. a device pixel).
http://www.w3.org/QA/Tips/font-size made good sense when written, continues
to be a best practice, and should be adhered to by every web leader of the
W3's stature.

[1] device pixels, the pixels most web users are stuck with, while classed as
relative in the CSS specs, aren't relevantly relative. This is because the
viewing device pixels are relative to is not a characteristic that the viewer
ordinarily can adjust to suit when the output is different from that
required, unlike the other CSS units classed as relative.
"Our Constitution was made only for a moral and religious
people. It is wholly inadequate to the government of any
other."                      John Adams, 2nd US President

 Team OS/2 ** Reg. Linux User #211409

Felix Miata  ***  http://fm.no-ip.com/
Received on Tuesday, 19 January 2010 15:33:22 UTC

This archive was generated by hypermail 2.4.0 : Monday, 18 April 2022 20:33:48 UTC