About font sizes

[deko]

Re: About font sizes

> Percentage also allows the author to respect the user's font size preference

That's not true if I use a percentage that specifies a non-existent font size.
We are back to the crap shoot. The only way a percentage has any correlation to
a user preference is if that percentage is 100%. Otherwise you never know what
you're going to get - that is not respecting the user's preferences.

If I specify 125% for h2, which should be about 20px, I can only hope there is a
corresponding font with a 20px definition. Maybe there is something close, but
maybe there isn't. For all I know, the browser may not change the font size at
all.

So it makes absolutely no sense to use percentages.

The only way to give the user control over font size is to not specify any.

The goal here is simply to allow the user to adjust the font size if he desires.

It makes no difference whatsoever if a web page defines font size in percentage
or px if the user's browser can scale the font size. If all the mainstream
browsers have the ability to scale px then you are better off using px to ensure
proper presentation of the page.

[Alan J. Flavell]

Re: About font sizes

On Sun, 11 Jun 2006, deko wrote:
[color=blue]
> If I specify 125% for h2, which should be about 20px, I can only
> hope there is a corresponding font with a 20px definition. Maybe
> there is something close, but maybe there isn't. For all I know,
> the browser may not change the font size at all.[/color]

Lynx won't change the font size. Neither will a speaking browser.

Presentation proposals are optional, by design.
[color=blue]
> So it makes absolutely no sense to use percentages.[/color]

By your argument, it makes no sense to propose /any/ font size, since
some browsers, by design, will ignore the proposal. I don't accept
that argument.
[color=blue]
> The only way to give the user control over font size is to not
> specify any.[/color]

Illogical, Mr. Spock. Most browsers which implement font sizing have
a reasonable range of font sizes available, and will use them. The
CSS specification includes some practical suggestions for usable size
increments.
[color=blue]
> It makes no difference whatsoever if a web page defines font size in
> percentage or px if the user's browser can scale the font size.[/color]

The specification allows for the possibility that the user could
/override/ the author's font size choices. Please be quite clear that
a specification-conforming override of absolute font sizing would not
scale the absolute size specifications - it would /override/ them.
[color=blue]
> If all the mainstream browsers have the ability to scale px[/color]

If you rule out browsers which implement the specification!
[color=blue]
> then you are better off using px to ensure proper presentation of
> the page.[/color]

Your conclusion rests on a false premise (&/or depends on the fact
that there are browsers which choose to ignore the specification in
this regard). That's no way to decide on a good policy.


--

Beware of negative easements.

[deko]

Re: About font sizes

>> If I specify 125% for h2, which should be about 20px, I can only[color=blue][color=green]
>> hope there is a corresponding font with a 20px definition. Maybe
>> there is something close, but maybe there isn't. For all I know,
>> the browser may not change the font size at all.[/color]
>
> Lynx won't change the font size. Neither will a speaking browser.
>
> Presentation proposals are optional, by design.
>[color=green]
>> So it makes absolutely no sense to use percentages.[/color]
>
> By your argument, it makes no sense to propose /any/ font size, since
> some browsers, by design, will ignore the proposal. I don't accept
> that argument.[/color]

Not that the proposal will be ignored, but rather that the proposal will be
misinterpreted.

Using a percentage-based scale that requires rounding to a valid font size is
inherently inaccurate if the basis for the scale (user preference) is unknown.
[color=blue][color=green]
>> The only way to give the user control over font size is to not
>> specify any.[/color]
>
> Illogical, Mr. Spock. Most browsers which implement font sizing have
> a reasonable range of font sizes available, and will use them. The
> CSS specification includes some practical suggestions for usable size
> increments.[/color]

This is meaningless unless we all agree on a basis for the scale. If everyone
who uses a browser agrees to set his default font size to 16, then we have a
chance at hitting those usable size increments, otherwise it's a crap shoot.
[color=blue][color=green]
>> It makes no difference whatsoever if a web page defines font size in
>> percentage or px if the user's browser can scale the font size.[/color]
>
> The specification allows for the possibility that the user could
> /override/ the author's font size choices. Please be quite clear that
> a specification-conforming override of absolute font sizing would not
> scale the absolute size specifications - it would /override/ them.[/color]

well, this is semantics... override vs. scale... who cares. turn the knob one
way, the font gets bigger; the other way, smaller. That's all that matters.

[Alan J. Flavell]

Re: About font sizes

On Sun, 11 Jun 2006, deko wrote:
[color=blue][color=green]
> > By your argument, it makes no sense to propose /any/ font size,
> > since some browsers, by design, will ignore the proposal. I don't
> > accept that argument.[/color]
>
> Not that the proposal will be ignored, but rather that the proposal
> will be misinterpreted.[/color]

But that's exactly what you're hoping for! You say you want to
specify pt or px units, *but* you want the browser to misinterpret
them as relative units. That isn't what the specification says, and
so you're relying on the generosity of browsers like Moz, Opera etc.
to disregard what you're asking for, and treat it as something
different. Whereas IE, in this respect, is doing what the
specification calls for, and taking you at your word that absolute
sizing is required.

If, on the other hand, you would use appropriate sizing units, you
wouldn't need to rely on user-friendly browsers not taking your
proposals seriously, and you wouldn't have to face the problem of
visually disabled users of IE needing to find an obscure option to
disable font sizing.
[color=blue]
> Using a percentage-based scale that requires rounding to a valid
> font size is inherently inaccurate if the basis for the scale (user
> preference) is unknown.[/color]

That's the reality of designing for the web. So cope with it.
[color=blue][color=green]
> > Most browsers which implement font sizing have a reasonable range
> > of font sizes available, and will use them. The CSS specification
> > includes some practical suggestions for usable size increments.[/color]
>
> This is meaningless unless we all agree on a basis for the scale.[/color]

How so? It's not ideal, but it adapts better to a wide range of
presentation situations than anything else I have seen. If you were
so determined to get pixel-exact rendering, you'd be better advised to
use PDF - but I don't suppose your users would thank you for it.
[color=blue]
> If everyone who uses a browser agrees to set his default font size
> to 16,[/color]

What colour is the sky on your planet?
[color=blue][color=green]
> > The specification allows for the possibility that the user could
> > /override/ the author's font size choices. Please be quite clear
> > that a specification-conforming override of absolute font sizing
> > would not scale the absolute size specifications - it would
> > /override/ them.[/color]
>
> well, this is semantics...[/color]

Seems to me that you're missing the point.
[color=blue]
> override vs. scale... who cares.[/color]

If the user overrides your sizing, then your sizing proposals have
*no* effect. The user gets h1, h2, p, big, small etc. sized as
determined by /their/ browser or stylesheet, irrespective of your
intentions.
[color=blue]
> turn the knob one way, the font gets bigger; the other way, smaller.
> That's all that matters.[/color]

Really?

[Alan J. Flavell]

Re: About font sizes

On Sat, 10 Jun 2006, Jukka K. Korpela wrote:
[color=blue]
> Typography rules say, among other things, that a change in font size
> should be clear enough so that it does not look like an error. The
> mileage varies, but let us assume that a 10% increase is desired. So
> if we say font-size: 110%, what will happen? The browser computers
> the absolute font size and then selects the closest actual font size
> it can use - at the simplest, selecting among differently sized
> fonts for the given font family (typeface). For all we can know,
> that size could be identical to the basic font size (when only a few
> sizes exist). Or it might be almost 20% larger than the basic font
> size.[/color]

I don't disagree with what you have written, as a theory. But what
conclusions do you draw in terms of recommendations to authors?
[color=blue]
> For example, if the basic font size is 12pt, then 110% yields
> 13.2pt, which I would expect to get rounded to 13pt. However, for
> Times New Roman, I seem get a bigger font (13.5pt?), which is rather
> big and looks bolded. If I set the font size to a smaller
> percentage, there's the risk of getting no font size increase, for
> some combinations of font face and basic size.[/color]

Yes; and I get very different visual results depending on how I have
set my font smoothing options. But these details of user settings are
not something which an author can (or should) know, nor take into
account in their CSS.

regards

[deko]

Re: About font sizes

>> > By your argument, it makes no sense to propose /any/ font size,[color=blue][color=green][color=darkred]
>> > since some browsers, by design, will ignore the proposal. I don't
>> > accept that argument.[/color]
>>
>> Not that the proposal will be ignored, but rather that the proposal
>> will be misinterpreted.[/color]
>
> But that's exactly what you're hoping for! You say you want to
> specify pt or px units, *but* you want the browser to misinterpret
> them as relative units.[/color]

Let's assume for a moment a browser has all defaults, no user settings, no
overrides. If I specify 14px, then the browser will render 14px. Where is the
misinterpretati on there?

If the user has defined preferences, however, then *whatever* I've specified,
*regardless* if I used px of percentage, will be interpreted (or misinterpreted,
as the case may be).
[color=blue]
> That isn't what the specification says, and
> so you're relying on the generosity of browsers like Moz, Opera etc.
> to disregard what you're asking for, and treat it as something
> different. Whereas IE, in this respect, is doing what the
> specification calls for, and taking you at your word that absolute
> sizing is required.
>
> If, on the other hand, you would use appropriate sizing units[/color]

Appropriate? The only thing that's appropriate is a font size the browser can
display. What difference does it make if I say 100% or 16px? What's displayed
in the client's browser is a matter of preference - known only by the user.
[color=blue]
> you
> wouldn't need to rely on user-friendly browsers[/color]

There is absolutely no reason *not* to rely on such browser features. Thinking
otherwise is a completely Luddite attitude.
[color=blue]
> not taking your
> proposals seriously, and you wouldn't have to face the problem of
> visually disabled users of IE needing to find an obscure option to
> disable font sizing.[/color]

Fair enough. Visually disabled users will hate my site.
[color=blue][color=green]
>> Using a percentage-based scale that requires rounding to a valid
>> font size is inherently inaccurate if the basis for the scale (user
>> preference) is unknown.[/color]
>
> That's the reality of designing for the web. So cope with it.
>[color=green][color=darkred]
>> > Most browsers which implement font sizing have a reasonable range
>> > of font sizes available, and will use them. The CSS specification
>> > includes some practical suggestions for usable size increments.[/color]
>>
>> This is meaningless unless we all agree on a basis for the scale.[/color]
>
> How so? It's not ideal, but it adapts better to a wide range of
> presentation situations than anything else I have seen. If you were
> so determined to get pixel-exact rendering, you'd be better advised to
> use PDF - but I don't suppose your users would thank you for it.
>[color=green]
>> If everyone who uses a browser agrees to set his default font size
>> to 16,[/color]
>
> What colour is the sky on your planet?
>[color=green][color=darkred]
>> > The specification allows for the possibility that the user could
>> > /override/ the author's font size choices. Please be quite clear
>> > that a specification-conforming override of absolute font sizing
>> > would not scale the absolute size specifications - it would
>> > /override/ them.[/color]
>>
>> well, this is semantics...[/color]
>
> Seems to me that you're missing the point.[/color]

What point? The only point is this:

*Users should be able to adjust the size of the text in web site.*

I'm suggesting that they can do this regardless of how the page author has
specified font sizes in his markup.
[color=blue][color=green]
>> override vs. scale... who cares.[/color]
>
> If the user overrides your sizing, then your sizing proposals have
> *no* effect. The user gets h1, h2, p, big, small etc. sized as
> determined by /their/ browser or stylesheet, irrespective of your
> intentions.[/color]

And would this be any different if I used percentage sized fonts? No, it would
not make any difference at all.
[color=blue][color=green]
>> turn the knob one way, the font gets bigger; the other way, smaller.
>> That's all that matters.[/color]
>
> Really?[/color]

Yes, of course.

[Jim Moe]

Re: About font sizes

deko wrote:[color=blue]
> As I understand it, most browser manufacturers have agreed on 16px for their
> default font size.
>
> So, this should be an accurate conversion for percentages:
> px %
> 16 = 100
> 14 = 87.5
> 13 = 81.25
> 12 = 75
> 10 = 62.5
>[/color]
Yes. And it is basically useless as a layout guide.
Use percentages that produce the type of result you wish to display for
a variety of default font sizes.
[color=blue]
> I assume it's better to stick to a percentage that will yield a round number -
> rather than using something like 90% to get 14.4. A browser will just round up
> or down to a font size it can display, correct?
>[/color]
The number of decimal places in a percentage value is irrelevant. Round
numbers are only easier for humans, not machines.
[color=blue]
> As for em, am I correct in saying 1em = 16px?
>[/color]
No. It is the distance between baselines with a default line-height.
Although for practical purposes it is close enough.
[color=blue]
> Does that mean 1.2em is 19.2px (rounded by the browser to 19px - if the default
> size is 16)?
>[/color]
Only if, in the current CSS context, the font size corresponds to 16px.
If the font size has changed to font-size:125% (and assuming the original
font-size was 16px), 1.2em is then 1.2 * (125% * 16px) = 24px.

body { font-size: 100%; }
p { font-size: 125%; }

--
jmm (hyphen) list (at) sohnen-moe (dot) com
(Remove .AXSPAMGN for email)

[Jukka K. Korpela]

Re: About font sizes

Jim Moe <jmm-list.AXSPAMGN@s ohnen-moe.com> scripsit:
[color=blue][color=green]
>> As for em, am I correct in saying 1em = 16px?
>>[/color]
> No. It is the distance between baselines with a default line-height.[/color]

No, 1em is the size of the font, and for font-size, it exceptionally means
the font size of the enclosing element ("parent").

The distance between baselines is, by definition, the line height, which is
generally larger than 1em. Typical browser defaults for line height are
around 1.2em.
[color=blue]
> Although for practical purposes it is close enough.[/color]

For practical purposes, 1em can be anything when mapped to a physical size.
I don't think nobody argues about the observation that 16px is the most
common browser default.

--
Jukka K. Korpela ("Yucca")

http://www.cs.tut.fi/~jkorpela/

[Jukka K. Korpela]

Re: About font sizes

Alan J. Flavell <flavell@physic s.gla.ac.uk> scripsit:
[color=blue]
> I don't disagree with what you have written, as a theory. But what
> conclusions do you draw in terms of recommendations to authors?[/color]

I really don't know. It seems that one cannot use percentages reliably to
create suitable variation in font size in text. Using the size keywords
(e.g., font-size: larger) doesn't work well either, since I cannot know what
they are really mapped to.
[color=blue]
> But these details of user settings are
> not something which an author can (or should) know, nor take into
> account in their CSS.[/color]

The point is that this situation makes font size variation inside a document
rather dubious, except for headings and other texts for which we use rather
large font sizes. For example, to indicate a passage as less important, any
method for making its font size smaller seems to be rather risky - the main
risks being that the result is too small and that there is no actual change
in font size

--
Jukka K. Korpela ("Yucca")

http://www.cs.tut.fi/~jkorpela/

[Darin McGrew]

Re: About font sizes

deko <deko@nospam.co m> wrote:[color=blue]
> Using a percentage-based scale that requires rounding to a valid font size is
> inherently inaccurate if the basis for the scale (user preference) is unknown.[/color]

Does it really matter whether my default font size is 13px and your heading
(with a font-size of 130%) is 17px, or my default font size is 16px and
your heading is 22px?

If so, then HTML+CSS is the wrong medium for your content.
[color=blue]
> If everyone who uses a browser agrees to set his default font size to 16,[/color]

The phrase "unclear on the concept" comes to mind...
[color=blue]
> well, this is semantics... override vs. scale... who cares.[/color]

Well, here's a "scale" that you proposed using in an earlier message:

16px = 100%
14px = 87.5%
13px = 81.25%
12px = 75%
10px = 62.5%

and here's what my browser would display, as it's normally configured:

100% = 13px
87.5% = 12px
81.25% = 12px
75% = 12px
62.5% = 12px

So, what's the difference between the browser overriding the font size or
scaling the font size?
--
Darin McGrew, mcgrew@stanford alumni.org, http://www.rahul.net/mcgrew/
Web Design Group, darin@htmlhelp. com, http://www.HTMLHelp.com/

"Good teachers are costly. Bad teachers cost more." - Bob Talbert

[Alan J. Flavell]

Re: About font sizes

On Mon, 12 Jun 2006, Darin McGrew wrote:

[...][color=blue]
> So, what's the difference between the browser overriding the font
> size or scaling the font size?[/color]

The difference, as I see it, is between - on the one hand - an author
asking for something inappropriate (in this case, absolute or px size
units for use in the general web context), and relying on (most)
browsers to do something more sensible instead of what he asked for;
and - on the other hand - an author who is working *with* the medium,
rather than against it.

[deko]

Re: About font sizes

> Does it really matter whether my default font size is 13px and your heading[color=blue]
> (with a font-size of 130%) is 17px, or my default font size is 16px and
> your heading is 22px?[/color]

If it does not matter, then authors should not specify any font size at all!
What you are saying, in effect, is that users should program their browsers and
ignore the author's design proposal. That's silly. When you buy a magazine you
expect an attractive format by default. Would you want to program the magazine
with your explicit preferences through an interface on the cover before your
read it? Of course not.
[color=blue]
> Well, here's a "scale" that you proposed using in an earlier message:
>
> 16px = 100%
> 14px = 87.5%
> 13px = 81.25%
> 12px = 75%
> 10px = 62.5%
>
> and here's what my browser would display, as it's normally configured:
>
> 100% = 13px
> 87.5% = 12px
> 81.25% = 12px
> 75% = 12px
> 62.5% = 12px
>
> So, what's the difference between the browser overriding the font size or
> scaling the font size?[/color]

What you're saying is that if you visit a site with px-designated font sizes,
the fact that you can adjust the font size up or down is of little consequence
because the page does not display according to your preferences to begin with
(because your preferences are expressed in percentage while the author's
preferences are expressed in px). Fair enough. That's a reasonable argument.
But it's weakened by the fact that the vast majority of users would rather
accept an author's design than specify their own. Furthermore, it assumes every
page author is proposing an undesirable design.

This is perhaps the crux of the matter.

An undesirable design is one that is illegible. Thus it is incumbent upon an
author to design legible pages using reasonable font sizes. In the case of a
poorly designed page, or users with poor vision, the problem of legibility is
solved (for px-authored pages) by the browser's override feature which allows
users to increase or decrease the text size. Now if you say, "I will not
tolerate any author's grubby design! I must have complete control over every
piece of text my browser displays!", then we are no longer talking about
usability, or accommodating users with poor vision. We might as well argue
about the color of the rug.

I'm inclined to trust an author to provide a desirable design. This may be
because I have confidence in my own ability as a web designer. I agree
usability is paramount, but that does not mean using percentage-based font
sizes. In the absence of agreement on a size scale, px will ensure that a
design is rendered properly. If there were agreement, there would not be a
problem. When I suggested 16px as a default, I was only illustrating how this
would make things easier for everyone (by the way, all the major browser
manufactures did, in fact, agree, in 2000, to 16px for their browsers' default
font size setting).

When everyone does agree, we might have a more intuitive scale, perhaps:

xx-small
x-small
small
medium
large
x-large
xx-large

And, yes, it would be easier if government were a dictatorship ;)

[Darin McGrew]

Re: About font sizes

I wrote:[color=blue][color=green]
>> Does it really matter whether my default font size is 13px and your heading
>> (with a font-size of 130%) is 17px, or my default font size is 16px and
>> your heading is 22px?[/color][/color]

deko <deko@nospam.co m> wrote:[color=blue]
> If it does not matter, then authors should not specify any font size at all![/color]

Apparently I did not make myself clear.

Authors should leave the font size for body text alone. That is, they
should specify a font size of 100% for body text.

Headings and such can be specified with relative font sizes larger than
100%. Legalese and similar fine print that the average reader can safely
ignore can be specified with relative font sizes smaller than 100%.

If all you want is for the headings to be larger than the body text (e.g.,
130%), then HTML+CSS is an appropriate medium

If you want the body text and headings to be specific sizes (e.g., 16px and
22px), then HTML+CSS is not an appropriate medium. As others have
suggested, PDF might be a better alternative in such situations.
[color=blue]
> What you are saying, in effect, is that users should program their
> browsers and ignore the author's design proposal.[/color]

If necessary, yes.

If the author's design adapts to my display environment, then I'll leave it
alone. If it doesn't, then I'll switch to my browser's user mode and the
author's design is stripped away, leaving the bare content.
[color=blue]
> That's silly. When you buy a magazine you expect an attractive format by
> default.[/color]

The web is not a dead-trees magazine.

See http://www.westciv.com/style_master/...oil/not_paper/
[color=blue][color=green]
>> Well, here's a "scale" that you proposed using in an earlier message:
>>
>> 16px = 100%
>> 14px = 87.5%
>> 13px = 81.25%
>> 12px = 75%
>> 10px = 62.5%
>>
>> and here's what my browser would display, as it's normally configured:
>>
>> 100% = 13px
>> 87.5% = 12px
>> 81.25% = 12px
>> 75% = 12px
>> 62.5% = 12px
>>
>> So, what's the difference between the browser overriding the font size or
>> scaling the font size?[/color][/color]
[color=blue]
> What you're saying is that if you visit a site with px-designated font
> sizes, the fact that you can adjust the font size up or down is of little
> consequence because the page does not display according to your
> preferences to begin with (because your preferences are expressed in
> percentage while the author's preferences are expressed in px).[/color]

No. I'm saying that my browser will enforce a minimum font size. The
author's font sizes (whether specified in px, percentages, or something
else) will scale as long as they don't go below my minimum font size. If
they go below my minimum font size, then my browser will override them.
[color=blue]
> But it's weakened by the fact that the vast majority of users would rather
> accept an author's design than specify their own.[/color]

I admit that I am atypical. I configure my browser to display content in
spite of the author's broken design. Most users would just leave the site.
[color=blue]
> Furthermore, it assumes every page author is proposing an undesirable
> design.[/color]

Not at all. I simply configure my browser to enforce a few reasonable
constraints. Sites that adapt to my browsing environment are displayed in
all their glory. Sites that fight my browsing environment are displayed in
user mode, with the author's design stripped away.
[color=blue]
> In the absence of agreement on a size scale, px will ensure that a
> design is rendered properly.[/color]

No, it won't. For one thing, it creates the the illusion that you have
pixel-perfect control of your page layout.
--
Darin McGrew, mcgrew@stanford alumni.org, http://www.rahul.net/mcgrew/
Web Design Group, darin@htmlhelp. com, http://www.HTMLHelp.com/

Why is "abbreviate d" such a long word?

[Martin Eyles]

Re: About font sizes

"Beauregard T. Shagnasty" <a.nony.mous@ex ample.invalid> wrote in message
news:MWCig.1620 95$Fs1.74615@bg tnsc05-news.ops.worldn et.att.net...[color=blue]
> Chris Hughes wrote:
>[color=green]
>> 16px? That's eNORMous! (I just tried it.) 12px would be more
>> reasonable, surely?[/color]
>
> It would depend entirely on your visual acuity. Yours may be better than
> mine.
>
> A friend of mine has a 20 or 21" monitor set to 800x600 resolution so
> she can read it. Me, I can read it from the next room.
>
> --
> -bts
> -Warning: I brake for lawn deer[/color]

Why not have a high resolution, and large text icons etc. Big smooth text is
probably more readable than big blocky text

ME

[dingbat@codesmiths.com]

Re: About font sizes

deko wrote:
[color=blue]
> If it does not matter, then authors should not specify any font size at all![/color]

By George, I think he's got it!
[color=blue]
> What you are saying, in effect, is that users should program their browsers and
> ignore the author's design proposal.[/color]

Not ignore - join the cascade instead.
[color=blue]
> When you buy a magazine you expect an attractive format by default.[/color]

Yes, and for 500-year-old printed paper that's the best I could hope
for.

Now we have the web, and the author's size is a _default_, not a fixed
final size. I can adjust it, according to my personal circumstances
(which of course the author can't know about). This is actually an
improvement!

[color=blue][color=green]
> > Well, here's a "scale" that you proposed using in an earlier message:
> >
> > 16px = 100%[/color][/color]

Works great on my desktop, bit useless on my phone though.