Screen resolutions vs. web site legibility

[Dave]

A friend of mine pointed out the other day that certain elements on my web
site are too small. But in most of what I publish, fonts are at default size
or smaller, and my images are easy to see. I am viewing at 800x600 right now
by the way, because that's what I can comfortably see.

Times New Roman default pitch is what - 12px?

With people increasing their screen resolutions, is it possible for "default
size to ever become too small in the future? Will we be forced to "upscale"
our web pages to make them legible on larger resolutions (but like
childrens' books for those of us on smaller resolutions)?

If you know what I'm referring to, please explain. Any information is
appreciated. Thanks!

--
=============== =============
- Dave

http://members.cox.net/grundage/

[Neal]

Re: Screen resolutions vs. web site legibility

On Wed, 25 Aug 2004 15:33:19 -0500, Dave <dave@yahoo.com > wrote:
[color=blue]
> A friend of mine pointed out the other day that certain elements on my
> web
> site are too small. But in most of what I publish, fonts are at default
> size
> or smaller, and my images are easy to see. I am viewing at 800x600 right
> now
> by the way, because that's what I can comfortably see.
>
> Times New Roman default pitch is what - 12px?
>
> With people increasing their screen resolutions, is it possible for
> "default
> size to ever become too small in the future? Will we be forced to
> "upscale"
> our web pages to make them legible on larger resolutions (but like
> childrens' books for those of us on smaller resolutions)?
>
> If you know what I'm referring to, please explain. Any information is
> appreciated. Thanks!
>[/color]


The user is responsible for their user agent. If it is not delivering
content in an appropriate manner, the user must correct it in the browser.

Now, AS WE ALL KNOW BY NOW font-size should be set, if at all, in
percentages. 100% equals the user's default size, which has either been
set to what's appropriate, or wasn't all that bad to begin with. So, if
what you say is true, you are sending text with font-size: 100% (or no
font-size alterations at all) and the size seems incorrect on the screen,
then the visitor is at fault, and likely the user finds this problem with
a lot of content on a lot of sites. They should adjust the browser text
size to their preferred size.

However, if, as I suspect, you use font-size: 12px then you are not
sending text at default size. That size varies from screen to browser to
user. The only reliable way is to use 100%. Larger for headings and such,
no smaller *at all* unless the content doesn't need to be read by the
visitor.

How's that?

[Darin McGrew]

Re: Screen resolutions vs. web site legibility

Dave <dave@yahoo.com > wrote:[color=blue]
> A friend of mine pointed out the other day that certain elements on my web
> site are too small.[/color]

Well, the URL in your .sig <http://members.cox.net/grundage/> uses px for
font sizes. Many of those font sizes are smaller than the minimum font size
enforced by my browser, and fonts sized in px (and pt) units are
inappropriate on the WWW (see http://css.nu/faq/ciwas-aFAQ.html#QA02).

Instead, leave the font size alone for normal body text, and specify
exceptions (e.g., larger fonts for headings, smaller fonts for legalese) as
a percentage of the normal font size. You could also specify these
exceptions using em units, but then you need to include

body { font-size: 100% }

as a workaround for a rather obnoxious bug in MSIE.
[color=blue]
> Times New Roman default pitch is what - 12px?[/color]

It could be anything, assuming a font named "Times New Roman" even exists.
[color=blue]
> With people increasing their screen resolutions, is it possible for "default
> size to ever become too small in the future?[/color]

The user's normal font size will be legible. Your "default font size" may
or may not be legible on anyone else's system.

As Todd Fahrner wrote:

The font size chosen by the user as a comfortable default (1 em)
provides more truly useful information about the rendering
environment than all the resolution-sniffing, window-querying,
"open-this-wide" logic you can throw at the problem.

See also:

http://www.htmlhelp.com/faq/html/design.html#screen-size

http://www.xs4all.nl/~sbpoley/webmatters/flexdesign.html

--
Darin McGrew, mcgrew@stanford alumni.org, http://www.rahul.net/mcgrew/
Web Design Group, darin@htmlhelp. com, http://www.HTMLHelp.com/

"The early bird gets the worm, but the second mouse gets the cheese."

[Harlan Messinger]

Re: Screen resolutions vs. web site legibility


"Dave" <dave@yahoo.com > wrote in message
news:iI6Xc.6039 5$wo.33707@okep read06...[color=blue]
> A friend of mine pointed out the other day that certain elements on my web
> site are too small. But in most of what I publish, fonts are at default[/color]
size[color=blue]
> or smaller, and my images are easy to see. I am viewing at 800x600 right[/color]
now[color=blue]
> by the way, because that's what I can comfortably see.
>
> Times New Roman default pitch is what - 12px?
>
> With people increasing their screen resolutions, is it possible for[/color]
"default[color=blue]
> size to ever become too small in the future? Will we be forced to[/color]
"upscale"[color=blue]
> our web pages to make them legible on larger resolutions (but like
> childrens' books for those of us on smaller resolutions)?[/color]

Browsers are programmed to take the screen resolution into account when
displaying content. Users can then adjust the default size to suit their own
preferences. There is no problem unless *you* define the font-size in terms
of pixels, points, or picas. Define them as ems or %s.

[Steve Pugh]

Re: Screen resolutions vs. web site legibility

"Dave" <dave@yahoo.com > wrote:
[color=blue]
>A friend of mine pointed out the other day that certain elements on my web
>site are too small. But in most of what I publish, fonts are at default size
>or smaller,[/color]

So some fonts are smaller than default and you wonder why some people
think they are too small?
[color=blue]
> and my images are easy to see. I am viewing at 800x600 right now
>by the way, because that's what I can comfortably see.
>
>Times New Roman default pitch is what - 12px?[/color]

12px, 14px, 16px or 20px depending on browser and OS.
16px for most users as that's the default in Windows IE.
But the user can alwsys configure a different default if they choose.
[color=blue]
>With people increasing their screen resolutions, is it possible for "default
>size to ever become too small in the future? Will we be forced to "upscale"
>our web pages to make them legible on larger resolutions (but like
>childrens' books for those of us on smaller resolutions)?[/color]

If someone opts to use a larger resolution but also opts not adjust
they default font size then they can really only blame themselves.

Don't specify font sizes in pixels.
Don't specify the main body copy at a smaller size than default.

Steve

--
"My theories appal you, my heresies outrage you,
I never answer letters and you don't like my tie." - The Doctor

Steve Pugh <steve@pugh.net > <http://steve.pugh.net/>

[Shawn K. Quinn]

Re: Screen resolutions vs. web site legibility

[crossposted to, and followups set for, c.i.w.a.stylesh eets]

Dave wrote:
[color=blue]
> A friend of mine pointed out the other day that certain elements on my web
> site are too small.[/color]

It's possible that his/her browser is configured incorrectly.
[color=blue]
> But in most of what I publish, fonts are at default size or smaller, and
> my images are easy to see. I am viewing at 800x600 right now by the way,
> because that's what I can comfortably see.
>
> Times New Roman default pitch is what - 12px?[/color]

Who says they have Times New Roman installed *at all*? "World Wide Web
browser" does not imply "PC running the latest Microsoft Windows".

And default font size can be anywhere between, say, 6px and 60px, sometimes
even outside those ranges.
[color=blue]
> With people increasing their screen resolutions, is it possible for
> "default size to ever become too small in the future?[/color]

The size of 1em will never become too small in the future. Anything else,
especially based in something like px units, makes assumptions about the
user's system that will likely be horribly wrong at some point.

--
Shawn K. Quinn

[Brian]

Re: Screen resolutions vs. web site legibility

Dave wrote:
[color=blue]
> in most of what I publish, fonts are at default size
> or smaller, and my images are easy to see. I am viewing at 800x600 right now
> by the way, because that's what I can comfortably see.
>
> Times New Roman default pitch is what - 12px?[/color]

I see how much time you spent reading the group before jumping in. If
you had spent some time, you might have come across this:

<quote>
The font size chosen by the user as a comfortable default (1 em)
provides more truly useful information about the rendering environment
than all the resolution-sniffing, window-querying, "open-this-wide"
logic you can throw at the problem.
</quote>
[color=blue]
> With people increasing their screen resolutions, is it possible for "default
> size to ever become too small in the future?[/color]

Only if you don't read the group and follow the advice so often suggested.
[color=blue]
> Will we be forced to "upscale"
> our web pages to make them legible on larger resolutions (but like
> childrens' books for those of us on smaller resolutions)?[/color]

Try font-size: 100%; (Note that MSIE/Win is buggy when fonts are set in
em units. Percentages are equivalent, and help work around the bugs.)
100% sets the font size to 100% of the size selected in the user's browser.

--
Brian (remove ".invalid" to email me)

http://www.tsmchughs.com/

[Chris Morris]

Re: Screen resolutions vs. web site legibility

Neal <neal413@yahoo. com> writes:[color=blue]
> Now, AS WE ALL KNOW BY NOW font-size should be set, if at all, in
> percentages. 100% equals the user's default size, which has either
> been set to what's appropriate, or wasn't all that bad to begin
> with. So, if what you say is true, you are sending text with
> font-size: 100% (or no font-size alterations at all) and the size
> seems incorrect on the screen, then the visitor is at fault, and
> likely the user finds this problem with a lot of content on a lot of
> sites. They should adjust the browser text size to their preferred
> size.[/color]

I've been thinking about this issue a bit.

The *vast* majority of sites on the web display text at a size below
the browser default size. And this is wrong and were we starting again
with the web we wouldn't do this. But it is the case now.

Ignore the sites that set their size in CSS pt/px which completely
ignores the browser default size, because there's nothing that can be
done about them apart from the minimum size setting that's appeared in
modern browsers, and just consider the ones that use em, %, or the old
<font size="2"> - in other words, sizes relative to the browser
default.

If I set my browser text size to the size that I find comfortable for
reading body text, then the vast majority of pages on the web will
appear too small [1]. If I set my browser text size to the size that I
find comfortable by going to a typical web page and adjusting the size
until its comfortable, I will have a default size that is actually
somewhat larger than my preferred size, but does have the advantage of
working on the majority of web sites I visit.

Neilsen (in http://useit.com/alertbox/991114.html) reckons that if 80%
of the major sites on the web make a certain design error [2], then it's
often easier on users to make the same error and have consistent
behaviour then to go for something technically better but likely to be
unfamiliar to users. Unless it's a *lot* better.

So:

If I set my _site's_ font size to 100%/1em, then those people who have
configured their browsers to display the majority of websites at a
comfortable size will find the text too large.

If I set my site's font size to slightly less than this (85%,
perhaps), then those people who have configured their browsers to
display the majority of websites at a comfortable size will find the
text also at their preferred size (give or take a point or so). People
who find this font size too small are also likely to find the vast
majority of websites too small (and will hopefully do something about
it).

There is the argument that it's better to have text too large than too
small. However, the RNIB site (http://www.rnib.org.uk/) appears to
have a base text size of 80% for most elements, and that site was
heavily tested with partially-sighted users.

To get to the point - are we right to be recommending 100% as the base
font size? Yes, in principle we are. In practice I'm not so
sure. Provided relative units are used, something a little smaller
than 100% appears to be safe in practice.

So, given that we always recommend 1em on this and other newsgroups,
and we must have very good reasons for doing this, why am I wrong -
why is 100% *in practice* the right size to use?

[1] Unless I also set my minimum size to the same one, but if I do
that it doesn't really matter *what* the page sets as font size, I
won't notice at all.

[2] Now, most websites don't have alt attributes, but they can be
added in without anyone who doesn't need them noticing, so that's not
the same sort of issue.

--
Chris

[Alan J. Flavell]

Re: Screen resolutions vs. web site legibility

On Thu, 26 Aug 2004, Chris Morris wrote:

[...][color=blue]
> To get to the point - are we right to be recommending 100% as the base
> font size? Yes, in principle we are.[/color]

Right.
[color=blue]
> In practice I'm not so sure. Provided relative units are used,
> something a little smaller than 100% appears to be safe in practice.[/color]

As you say, there's an argument in favour of making the same mistakes
as the majority of web authors make. So all link texts should read
"click here" ? No thanks. External links should consistently open in
a new window ? No thanks.
[color=blue]
> So, given that we always recommend 1em on this and other newsgroups,
> and we must have very good reasons for doing this,[/color]

We are, to some extent, idealists. We don't like it when theory and
practice are out of alignment, and we try to persuade people to bring
them more closely into line. In the end, they'll do whatever they're
going to do... I don't really think they need any extra encouragement
from us.
[color=blue]
> why am I wrong -
> why is 100% *in practice* the right size to use?[/color]

You're sort of proposing that the advice they get from us would be "do
it this way - it's wrong in principle, but most other authors do it
this (wrong) way, and it's what readers have grown to expect" - at
least, that's how I interpret what you're saying.

I'm unhappy with that, I have to say.

[Chris Morris]

Re: Screen resolutions vs. web site legibility

"Alan J. Flavell" <flavell@ph.gla .ac.uk> writes:[color=blue]
> On Thu, 26 Aug 2004, Chris Morris wrote:
> [...][color=green]
> > To get to the point - are we right to be recommending 100% as the base
> > font size? Yes, in principle we are.[/color]
>
> Right.
>[color=green]
> > In practice I'm not so sure. Provided relative units are used,
> > something a little smaller than 100% appears to be safe in practice.[/color]
>
> As you say, there's an argument in favour of making the same mistakes
> as the majority of web authors make. So all link texts should read
> "click here" ? No thanks. External links should consistently open in
> a new window ? No thanks.[/color]

Well, Neilsen's caveat was that if it gave a large benefit to do it
the right way, it was still a good idea. In those cases the benefit is
clearly large enough (and the accessibility issue is much larger as
well). In the case of fonts, does it generally give that big an
improvement to make the font size larger? (Fitting less information on
the screen, fewer words/line at narrow window widths, etc mean that
larger font sizes can be to some extent harmful in some cases, as
well)
[color=blue][color=green]
> > So, given that we always recommend 1em on this and other newsgroups,
> > and we must have very good reasons for doing this,[/color]
>
> We are, to some extent, idealists. We don't like it when theory and
> practice are out of alignment, and we try to persuade people to bring
> them more closely into line. In the end, they'll do whatever they're
> going to do... I don't really think they need any extra encouragement
> from us.[/color]

True. But in *theory*, blue is an awful unvisited link colour,
especially as paired with red/purple as the visited - it's incredibly
tricky to get colours that are both nicely readable and have the blue
as the more prominent colour. Certainly not all browser defaults
manage it.

In practice we recommend blue for unvisited links because of user
familiarity, but if we were starting again with the web we'd use
something else. Likewise if we were starting again we'd have 1em as
the font size, but in practice doing something theoretically
non-optimal may be better for users.

(Some of our intranet pages have red for unvisited and a purple for
visited. In theory, this is a better colour set. In practice readers
are used to blue; I certainly get confused enough by these different
link colours)
[color=blue][color=green]
> > why am I wrong -
> > why is 100% *in practice* the right size to use?[/color]
>
> You're sort of proposing that the advice they get from us would be "do
> it this way - it's wrong in principle, but most other authors do it
> this (wrong) way, and it's what readers have grown to expect" - at
> least, that's how I interpret what you're saying.[/color]

That's about it, yes.
[color=blue]
> I'm unhappy with that, I have to say.[/color]

I don't particularly like it, no, as a principle. But on the other
hand, the *aim* of setting 100% is to get the reader's preferred font
size. And the majority of sites use something closer to 80%. So the
preferred font size is likely to be set to cope with that (if it's
been set [2]) and so 100% will actually be ~1/0.8 of their preferred
font size - which isn't what we're _trying_ to set the size to.

It does feel completely wrong to advocate setting font size smaller
than the user's preferred size so that the text will be approximately
the user's preferred size. But on the other hand the alternative is to
advise that people use font sizes that are in practice too large for
most users (I've had complaints from users that 90% was too large,
none that it was too small [3]).

With the ability in modern browsers to set minimum font sizes, the
issue will probably get less important in future anyway, but it's
going to be some time before the market share is high enough.

[2] If it's not been, it's either fine anyway or they have trouble
with nearly every site on the web, in which case an extra 15-25% font
size probably won't be a lot of use.
[3] Though of course, they may not have been able to find the
complaints form because the link was too small...

--
Chris

[Philipp Lenssen]

Re: Screen resolutions vs. web site legibility

Dave wrote:
[color=blue]
> A friend of mine pointed out the other day that certain elements on
> my web site are too small. But in most of what I publish, fonts are
> at default size or smaller, and my images are easy to see. I am
> viewing at 800x600 right now by the way, because that's what I can
> comfortably see.
>
> Times New Roman default pitch is what - 12px?
>[/color]

Just stop using any font-sizes, if in doubt. This will always fall-back
to the user's default font-size setting, which should be OK in 99% of
the cases (and in 1% can be blamed on misconfiguratio n on the user's
side). For footnotes, you might also use something like font-size: 90%
in the CSS.

--
Google Blogoscoped

http://blog.outer-court.com

[Neal]

Re: Screen resolutions vs. web site legibility

On 26 Aug 2004 12:22:22 +0100, Chris Morris <c.i.morris@dur ham.ac.uk>
wrote:
[color=blue]
> In the case of fonts, does it generally give that big an
> improvement to make the font size larger? (Fitting less information on
> the screen, fewer words/line at narrow window widths, etc mean that
> larger font sizes can be to some extent harmful in some cases, as
> well)[/color]

It's not a matter of making the font larger, it's a matter of making it
manipulatable by the user.

When you set font-size in points or pixels (which may or may not be
smaller than the ser can read) or at % less than 100 (which is likely to
be smaller than the user can read), you're gaining more content above the
fold, but at the cost of taking power and utility away from the user! I
can't support that. I want my content read.

If we use 100% for font-size, each user can resize in the browser. If we
use other measurements, some to many users lose that functionality. The
user is the one who knows what size text is legible, not me, so the user
requires the power to resize. Using % is the only way to offer that power
to the user. Using less than 100% means those users who cannot read below
100% of their set font size preference must reset the browser for my site.

I believe sites fall into a few categories here. Most sites are still
using pt or px and IE is screwed on the resize thing, so there's no sense
in worrying about those sites, they will appear the same no matter what is
done. Some sites use only ems or small-size %, they can pose problems
here, but they are a minority. Then there are sites following the 100%
rule, or which do not alter font-size at all. These will resize as users
expect.

Unless I've neglected something, there seems to be at best a tiny loss in
ability to read sites at large when the default size is adjusted to what
the user prefers.
[color=blue]
> True. But in *theory*, blue is an awful unvisited link colour,
> especially as paired with red/purple as the visited - it's incredibly
> tricky to get colours that are both nicely readable and have the blue
> as the more prominent colour. Certainly not all browser defaults
> manage it.
>
> In practice we recommend blue for unvisited links because of user
> familiarity, but if we were starting again with the web we'd use
> something else. Likewise if we were starting again we'd have 1em as
> the font size, but in practice doing something theoretically
> non-optimal may be better for users.[/color]

The only two reasons I can think of for dramatically altering the color of
links is when there is insufficient contrast for readability, or the color
simply clashes with the design of the site. In either case a value close
enough to the usual color can be easily discovered which will work.

If we were starting over with the web - which, I must stress, we are not!
- we'd do best to ensure readability of each document by stressing that
100% is the user's preference, below that may not be readable, above that
is good for headings and such, and other measurements are not wise for web
use (but may be appropriate for intranets and other non-webpage use). Of
course, while we started over we'd make the browsers able to resize text
more precisely - maybe a type-in box in the settings where we can set
exactly the size we prefer for our default text.
[color=blue]
> (Some of our intranet pages have red for unvisited and a purple for
> visited. In theory, this is a better colour set. In practice readers
> are used to blue; I certainly get confused enough by these different
> link colours)[/color]

In an intranet, since it's a closed, controlled environment, there's the
opportunity to do stuff like that and have it work. Any user who has
trouble knows where your office is to ask for help. Plus it's a captive
audience - they must use this intranet site for their job. These are two
major differences from the WWW.
[color=blue]
> I don't particularly like it, no, as a principle. But on the other
> hand, the *aim* of setting 100% is to get the reader's preferred font
> size. And the majority of sites use something closer to 80%. So the
> preferred font size is likely to be set to cope with that (if it's
> been set [2]) and so 100% will actually be ~1/0.8 of their preferred
> font size - which isn't what we're _trying_ to set the size to.[/color]

Again, I can only say that the user has chosen that size, whether by
browser used or dettings altered. So I have to say I have little sympathy
for a set of users who wish the font was smaller.
[color=blue]
> It does feel completely wrong to advocate setting font size smaller
> than the user's preferred size so that the text will be approximately
> the user's preferred size. But on the other hand the alternative is to
> advise that people use font sizes that are in practice too large for
> most users (I've had complaints from users that 90% was too large,
> none that it was too small [3]).[/color]

I've said it before - I'd rather annoy some users with big text than lose
some users with small text.

Big text is not as much of an issue. The aesthetics may not be optimal,
but if the page works and users can use it, the quality of the content
will help the user decide if the larger size (which happens to be what
they have actively or passively selected!!) is worth reading. If the size
is too small, the quality of the content cannot be discerned.

Ultimately, the "deeziner" crowd seems to think the purpose of a website
is to offer an attractive visual design, which it is not. The purpose of a
website is to deliver content to the visitor. That must take precedence
over any visual design issue.

In order to guarantee the visitor can read the content, we must give them
ultimate power in determining font-size. If they choose not to use that
power, that's sad. It's so easily done from a visible menu on all major
browsers. Serving body text at any size other than 100% disrupts the
sensibility and simplicity built into the way the WWW already works.

[Brian]

Re: Screen resolutions vs. web site legibility

Chris Morris wrote:
[color=blue]
> Ignore the sites that set their size in CSS pt/px which completely
> ignores the browser default size, because there's nothing that can be
> done about them apart from the minimum size setting that's appeared
> in modern browsers,[/color]

But, to borrow your phrase, "the vast majority of sites" do not set
their font size in flexible units.
[color=blue]
> and just consider the ones that use em, %, or the old <font size="2">
> - in other words, sizes relative to the browser default.[/color]

Ok, I'll accept that for the sake of the argument.
[color=blue]
> If I set my browser text size to the size that I find comfortable for
> reading body text, then the vast majority of pages on the web will
> appear too small [1].[/color]

Really? Or do you mean, the vast majority of web sites that override the
user's chosen size, but use % or em to override it? Those aren't the
same thing. I don't think I'm being pedantic. I set my font-size set to
what I find comfortable, but I do not notice any difference whatsoever
on the www at large. I suspect that if I increase my preferred size, I
would not notice any improvement.
[color=blue]
> If I set my site's font size to slightly less than this (85%,
> perhaps),[/color]

"Perhaps" is the key word. 85% is merely a guess. Moreover, I doubt that
many people change that font size, so how do you know that most people
are happy with font-size: 85%?

No matter what you use for font size, it will be a guess, unless you use
100%. From a practical point of view -- not just for the sake of an
ideal -- I see no reason to use anything but 100%.

--
Brian (remove ".invalid" to email me)

http://www.tsmchughs.com/

[Darin McGrew]

Re: Screen resolutions vs. web site legibility

Neal <neal413@yahoo. com> wrote:[color=blue]
> In an intranet, since it's a closed, controlled environment, there's the
> opportunity to do stuff like that and have it work. Any user who has
> trouble knows where your office is to ask for help. Plus it's a captive
> audience - they must use this intranet site for their job. These are two
> major differences from the WWW.[/color]

In my experience, intranets aren't as controlled an environment as some
designers think. My previous employer found out the hard way that not all
employees had access to the "correct" OS/browser, and were unable to use
some new intraweb sites that were rolled out with much fanfare. One was
required for benefits open enrollment, and there were no paper forms or
backup processes; you *had* to use the new intraweb process. Benefits open
enrollment was a mess that year.

And poor usability on an intranet costs real employee productivity, in
addition to the intangible issues (e.g., morale). And poor accessibility
may be more of an issue on an intranet, anywhere that disability access
laws protect employees more strongly than they protect the general public.

If anything, good WWW design principles are even more important on an
intranet.
--
Darin McGrew, mcgrew@stanford alumni.org, http://www.rahul.net/mcgrew/
Web Design Group, darin@htmlhelp. com, http://www.HTMLHelp.com/

"There are three kinds of people: those who can count and those who can't."

[Chris Morris]

Re: Screen resolutions vs. web site legibility

Neal <neal413@yahoo. com> writes:[color=blue]
> On 26 Aug 2004 12:22:22 +0100, Chris Morris <c.i.morris@dur ham.ac.uk>
> wrote:[color=green]
> > In the case of fonts, does it generally give that big an
> > improvement to make the font size larger? (Fitting less information on
> > the screen, fewer words/line at narrow window widths, etc mean that
> > larger font sizes can be to some extent harmful in some cases, as
> > well)[/color]
>
> It's not a matter of making the font larger, it's a matter of making
> it manipulatable by the user.[/color]

Except that font-size: 90% is just as resizable as font-size:
100%. Just in case I wasn't clear earlier, I am *not* in any way
suggesting px/pt is a good unit for WWW font-sizing.
[color=blue]
> When you set font-size in points or pixels (which may or may not be
> smaller than the ser can read) or at % less than 100 (which is likely
> to be smaller than the user can read), you're gaining more content[/color]

Hold on. I don't think that's true. Quite a lot of sites set a font
size smaller than the browser default. I don't think small amounts
less than 100% (80% and above, say) are likely to be unreadable.
[color=blue]
> above the fold, but at the cost of taking power and utility away from
> the user! I can't support that. I want my content read.[/color]

font-size: 90% takes no power away from the user.
[color=blue]
> If we use 100% for font-size, each user can resize in the browser. If
> we use other measurements, some to many users lose that
> functionality. The user is the one who knows what size text is
> legible, not me, so the user requires the power to resize. Using % is
> the only way to offer that power to the user. Using less than 100%
> means those users who cannot read below 100% of their set font size
> preference must reset the browser for my site.[/color]

Hmm. I find it difficult to believe that users set their default font
size to the absolute minimum they can read. If they did, they would
have to do a lot of resetting their browser as they moved about the
web because lots of sites use a smaller size.
[color=blue]
> I believe sites fall into a few categories here. Most sites are still
> using pt or px and IE is screwed on the resize thing, so there's no
> sense in worrying about those sites, they will appear the same no
> matter what is done.[/color]

Agreed in IE - Opera and Mozilla, OTOH, can rescale text set in px/pt
as if it were the equivalent (for whatever is the equivalent for that
browser) em value. So this may play a part in it.

And it is worth noting that sites that scale this way tend to set
10-13 px, or about 65-85% of the out-of-the-box settings of common
browsers. I'll come back to this below.
[color=blue]
> Some sites use only ems or small-size %, they can pose problems
> here, but they are a minority.[/color]

Don't forget all the sites that set size using <font>. They're also
relative-size. And generally small.
[color=blue]
> Unless I've neglected something, there seems to be at best a tiny loss
> in ability to read sites at large when the default size is adjusted to
> what the user prefers.[/color]

Is the default size set in the browser the one the user prefers as is,
or the one that gets the right result for them with the sites they
view. I suspect the latter.
[color=blue][color=green]
> > I don't particularly like it, no, as a principle. But on the other
> > hand, the *aim* of setting 100% is to get the reader's preferred font
> > size. And the majority of sites use something closer to 80%. So the
> > preferred font size is likely to be set to cope with that (if it's
> > been set [2]) and so 100% will actually be ~1/0.8 of their preferred
> > font size - which isn't what we're _trying_ to set the size to.[/color]
>
> Again, I can only say that the user has chosen that size, whether by
> browser used or dettings altered. So I have to say I have little
> sympathy for a set of users who wish the font was smaller.[/color]

Okay. Now those are two separate cases.
Choice by browser used:

- Either they don't know how to set font size, or they don't need to
know because the browser default works for them. The majority of
pages on the web will be displayed at ~80% of their browser default
size [because of the px/pt to em ratio on an out-of-the-box
browser]. These users, probably a (large?) majority, have no idea
even that browser settings can be altered [1]. When they come across a
site that uses 100% font size, it gets displayed considerably
larger than normal.

[1] We see enough people asking on this newsgroup who apparently
didn't know this before they turned up, and these are web developers
with enough general technical knowledge to find a newsreader too. What
hope does the average user have?

Choice by setting alteration:

Older browsers, including IE:
- The resizing won't work on the majority of pages. The mean (and
quite probably median) size on those that do will be less than
their *set* preference. Probably therefore their preference is set
larger than their actual preferred size.

Newer browsers:
- These have a 'minimum font size' setting. I needn't worry about
setting font-size unreadably small because if the user has properly
configured their browser settings, I can't. Tiny minority of users,
though, at the moment.

Browsers that ignore page font size settings entirely:
- Don't need to worry about these. I could set font-size: 1% and they
wouldn't notice.
[color=blue][color=green]
> > It does feel completely wrong to advocate setting font size smaller
> > than the user's preferred size so that the text will be approximately
> > the user's preferred size. But on the other hand the alternative is to
> > advise that people use font sizes that are in practice too large for
> > most users (I've had complaints from users that 90% was too large,
> > none that it was too small [3]).[/color]
>
> I've said it before - I'd rather annoy some users with big text than
> lose some users with small text.[/color]

Oh, agreed. I'm not advocating making a huge reduction in size below
100%, though.
[color=blue]
> Big text is not as much of an issue. The aesthetics may not be
> optimal, but if the page works and users can use it, the quality of
> the content will help the user decide if the larger size (which
> happens to be what they have actively or passively selected!!) is
> worth reading. If the size is too small, the quality of the content
> cannot be discerned.[/color]

I'm not talking about large drops in font-size here. As I said
earlier, the RNIB (who have an obvious interest in making sure their
site is readable by people with visual disabilities) have picked 80%
as their base size. I assume if this was causing problems for their
readers, they wouldn't be doing it.
[color=blue]
> Ultimately, the "deeziner" crowd seems to think the purpose of a
> website is to offer an attractive visual design, which it is not. The
> purpose of a website is to deliver content to the visitor. That must
> take precedence over any visual design issue.[/color]

True. I'm not suggesting changes that would (IMO) prevent content
being delivered, and I'm not suggesting them with the aim of making
the visual design look nice.

The aim of setting font-size: 100% is so that the text displays at the
users preferred size. I would argue that the majority of users who do
set a font size will set it to the size that makes *the sites they
use* have their preferred font size, which means that they've probably
set their browser default font size larger than their preferred font
size.

And, incidentally, can't the visual design be in some sense part of
the content? It identifies which site you're on, provides
semi-subliminal content of the sort that makes providing alt
attributes for semi-decorative photographs tricky, etc.
[color=blue]
> In order to guarantee the visitor can read the content, we must give
> them ultimate power in determining font-size. If they choose not to
> use that power, that's sad. It's so easily done from a visible menu on
> all major browsers.[/color]

True. It *is* moderately concealed in Internet Explorer and Mozilla,
though. If you know that your browser can do text resizing, then it's
quite easy to find it in the menus. If you *don't* know that your
browser can do text resizing, it's not immediately obvious that it
can.
[color=blue]
> Serving body text at any size other than 100% disrupts the
> sensibility and simplicity built into the way the WWW already works.[/color]

Now, this I think is the point I'm trying to make - it's *not* the way
the WWW already works. A tiny minority of sites have font-size: 100%;
even fewer have that as their minimum size. I can't think of a major
site [1] that does even the first (feel free to point a few out).

The original sensibility is already disrupted, and has been for
years. The thing that fixes it will be the mass adoption of browsers
with minimum font size settings.

[1] I mean a major general interest site - I know w3c.org does, for
example (though they use font-size: small for quite a few bits)

--
Chris