on text size...

[Jim Higson]

Most sites seem to use smaller text smaller than the user's default for
their content. Not just badly designed sties - a lot of the very beautiful
pages on CSS Zen Garden have text at 80% or smaller. From what I can tell,
this is because browsers (IE especially) have the default text size set to
be quite large.

Now, if users are used to text being set to 80% in CSS, they will set their
default size to 120% or so, thus getting it back up to a sane size. I do
this: I like reading 10px text, but if I keep that as the default I have to
hit Ctrl-plus for every site, so I set my browser to use size 14.

When designing sites I like to respect the users' preferences, and have
always thought the main content should be the default size, but once users
have compensated for every other site using 60~90% for their default text
this is making the text on my sites look large and out of place. Plus there
are many IE users who don't realise they can change their default text size
and just wonder why my text is so large.

The "font-size:80%" phenomenon seems to be a larger problem among web
designers of following the letter of the spec (validating HTML etc) but not
in the spirit by actually using the tools as intended.

What to do?
Jim

 

[Roy Schestowitz]

__/ [Jim Higson] on Monday 12 September 2005 10:08 \__

Most sites seem to use smaller text smaller than the user's default for
their content. Not just badly designed sties - a lot of the very beautiful
pages on CSS Zen Garden have text at 80% or smaller. From what I can tell,
this is because browsers (IE especially) have the default text size set to
be quite large.

Now, if users are used to text being set to 80% in CSS, they will set
their default size to 120% or so, thus getting it back up to a sane size.
I do this: I like reading 10px text, but if I keep that as the default I
have to hit Ctrl-plus for every site, so I set my browser to use size 14.

When designing sites I like to respect the users' preferences, and have
always thought the main content should be the default size, but once users
have compensated for every other site using 60~90% for their default text
this is making the text on my sites look large and out of place. Plus
there are many IE users who don't realise they can change their default
text size and just wonder why my text is so large.

The "font-size:80%" phenomenon seems to be a larger problem among web
designers of following the letter of the spec (validating HTML etc) but
not in the spirit by actually using the tools as intended.

What to do?
Jim

Fortunately, I still have decent eyesight, so I added the following to
userContent.css in my Firefox profile:

html{ font-size: 0.90em; }

What to do in the general case? There are no easy answers. You cannot change
millions of sites overnight and you cannot change the source code and
recompile a browser that hundreds of millions of people use. Your question
in some sense is a complaint or a request for concurrence, which you
definitely will get from me. A friend of mine who is retired chose to use
Firefox primarily because of flexibility in terms of text size.

I am afraid that I have nothing more to say at the moment. Others will have
more to say, I have no doubt about it.

Roy

 

[Neredbojias]

With neither quill nor qualm, Jim Higson quothed:

Most sites seem to use smaller text smaller than the user's default for
their content. Not just badly designed sties - a lot of the very beautiful
pages on CSS Zen Garden have text at 80% or smaller. From what I can tell,
this is because browsers (IE especially) have the default text size set to
be quite large.

Now, if users are used to text being set to 80% in CSS, they will set their
default size to 120% or so, thus getting it back up to a sane size. I do
this: I like reading 10px text, but if I keep that as the default I have to
hit Ctrl-plus for every site, so I set my browser to use size 14.

When designing sites I like to respect the users' preferences, and have
always thought the main content should be the default size, but once users
have compensated for every other site using 60~90% for their default text
this is making the text on my sites look large and out of place. Plus there
are many IE users who don't realise they can change their default text size
and just wonder why my text is so large.

The "font-size:80%" phenomenon seems to be a larger problem among web
designers of following the letter of the spec (validating HTML etc) but not
in the spirit by actually using the tools as intended.

What to do?

Go out drinking. In lieu of that, if you're capable at javascript, you
can detect the actual size of text on users' browsers and adjust it
accordingly. (Javascript users only, of course.)

Example:

http://www.neredbojias.com/alpha/rextex.html

 

[Toby Inkster]

What to do?

Don't be part of the problem.

 

[Jim Higson]

With neither quill nor qualm, Jim Higson quothed:


Go out drinking. In lieu of that

Can't I do both?

, if you're capable at javascript, you
can detect the actual size of text on users' browsers and adjust it
accordingly. (Javascript users only, of course.)

Ok, I'd say I'm capable at Javascript.
Check out my pet project beta (in Gecko)
http://81.5.150.113/wysi

It'll be released proper pretty soon.

Example:

http://www.neredbojias.com/alpha/rextex.html

This is interesting, but I'm not certain what they're doing is what I want.
They're resizing the text to always only just fill the page without
scrolling. I don't think this'd work very well for my 700 word articles.

 

[Jim Higson]

Don't be part of the problem.

Good call. There seems to be an arms race at the moment over who can have
the smallest text, especially at sites to do with HTML design.

1) Text is a bit big, but small text is where it's at, yo
2) Reduce by 10%
3) Users increase text size to compensate
4) Goto 1

The problem is, if I trust my users to have setup their browser properly,
how many will have? And of those who have set their font size, how many
will have set it while looking at the typical site which scales it down?

At a quick count, about 60% of the CSS Zen layouts on the front page are too
small for me to read if I set the font-size to be how I like it at 100%.

Eg:
http://www.csszengarden.com/?cssfile=/176/176.css&page=0
http://www.csszengarden.com/?cssfile=/178/178.css&page=0

Now, these are not pages styled by amateurs. Their authors should be the
ones who understand accessibility and respecting user preferences more than
anyone. Is there any reason very small text would be considered to the
spirit of the web standards?

 

[Alan J. Flavell]

The problem is, if I trust my users to have setup their browser properly,
how many will have?

For the users of Billware, they -all- have it set up the right size, and
here's my logic. Those who made a choice, got the size that they wanted
(within a certain tolerance); those who didn't make a choice have -paid-
Bill to set the right size for them - who are we to argue? Surely a
company as large as MS have studied their customer base closely enough to
know what default text size to set for them? If they needed to be asked
at installation time, Windows would ask them to choose - it asks them
enough other questions before the installation is complete, after all.

For the users of other OSes, in my experience, initial text size can vary
a lot, according to what installation choice was made of display
"resolution" etc.[1]; there may well be a dpi calibration, but the
installation dialogs usually don't take the user through setting it, and
many don't choose to calibrate the display with it. Instead, if they're
dissatisfied with the text size, they typically configure the
applications, rather than the display system. That's been my observation,
anyway.

[1] this won't change when we're all using panels, bearing in mind that
the expensive panels have higher values of pixels per inch, and thus are
apt for higher dpi settings, even though they're the same physical size.
(IBM are already over 200 pixels per inch for top-end displays.)

And we should never forget those with impaired vision, who have to set
very non-standard sizes to be able to use the web at all. Are they, like
the rest of us, to be asked to set their text size 50% larger (or
whatever) than they wanted it, so that headstrong authors can pull it back
down again to what they *really* needed?

And of those who have set their font size, how many
will have set it while looking at the typical site which scales it down?

"Me, Sir", I use Mozilla's friendly minimum font size setting to defend
myself against unreasonably scaled-down fonts, and recommend anyone seen
in a similar position to do the same.

[But I have to remember to turn that off again when I'm asked to review
someone else's web site...]

At a quick count, about 60% of the CSS Zen layouts on the front page are too
small for me to read if I set the font-size to be how I like it at 100%.

Eg:
http://www.csszengarden.com/?cssfile=/176/176.css&page=0
http://www.csszengarden.com/?cssfile=/178/178.css&page=0

Now, these are not pages styled by amateurs.

That's true...

Their authors should be the ones who understand accessibility and
respecting user preferences more than anyone.

Oh, really? I'm sorry, but I have to disagree with this. The
csszengarden is a spectacular demonstration of what -can- be achieved with
CSS - but under what are practically laboratory conditions: it is in no
way a role model to be used, as it stands, for production use on the web.

Its contributors, like all of us, have their strengths and weaknesses, and
furthermore they are somewhat limited by the rules of the zengarden.

Given that very few of us are experts in all of graphic design,
typography, accessibility and web technologies, I'd risk saying that the
best web pages, by and large, are the product of co-operative work
between, at least, two people: a graphic designer with some understanding
of the web, and a techie with some sympathy for graphic design.

Is there any reason very small text would be considered to the
spirit of the web standards?

No.

 

[Spartanicus]

At a quick count, about 60% of the CSS Zen layouts on the front page are too
small for me to read if I set the font-size to be how I like it at 100%.

Eg:
http://www.csszengarden.com/?cssfile=/176/176.css&page=0
http://www.csszengarden.com/?cssfile=/178/178.css&page=0

Now, these are not pages styled by amateurs. Their authors should be the
ones who understand accessibility and respecting user preferences more than
anyone.

Another reason why CSSZengarden is setting such a bad example. The
authors of the demos are *designers*, and designers understand and/or
care very little about/for accessibility. It's all about how it looks
for designers.

 

[Jim Higson]

For the users of Billware, they -all- have it set up the right size, and
here's my logic. Those who made a choice, got the size that they wanted
(within a certain tolerance); those who didn't make a choice have -paid-
Bill to set the right size for them - who are we to argue? Surely a
company as large as MS have studied their customer base closely enough to
know what default text size to set for them? If they needed to be asked
at installation time, Windows would ask them to choose - it asks them
enough other questions before the installation is complete, after all.

This is a good point, except of course if Bill noticed most sites use
smaller fonts and adjusts accordingly. :S

Anyway, the group has persuaded me to give my main text the default size
now.

For the users of other OSes, in my experience, initial text size can vary
a lot, according to what installation choice was made of display
"resolution" etc.[1]; there may well be a dpi calibration, but the
installation dialogs usually don't take the user through setting it, and
many don't choose to calibrate the display with it. Instead, if they're
dissatisfied with the text size, they typically configure the
applications, rather than the display system. That's been my observation,
anyway.

[1] this won't change when we're all using panels, bearing in mind that
the expensive panels have higher values of pixels per inch, and thus are
apt for higher dpi settings, even though they're the same physical size.
(IBM are already over 200 pixels per inch for top-end displays.)

Actually, I find this only true for laptops. I have a pretty high-end panel,
costing about £500 a while back, but it is only 1280*1024 at 19 inches -
the same number of pixels as a typical 17 incher. I've just checked a few
shops and most 19 inch desktop panels are still using this res.

If you know of a desktop TFT, 19 inches or bigger that does very high pixels
per inch please let me know. It's about time someone put a laptop
resolution panel a desktop casing.

And we should never forget those with impaired vision, who have to set
very non-standard sizes to be able to use the web at all. Are they, like
the rest of us, to be asked to set their text size 50% larger (or
whatever) than they wanted it, so that headstrong authors can pull it back
down again to what they *really* needed?

And of those who have set their font size, how many
will have set it while looking at the typical site which scales it down?

"Me, Sir", I use Mozilla's friendly minimum font size setting to defend
myself against unreasonably scaled-down fonts, and recommend anyone seen
in a similar position to do the same.

[But I have to remember to turn that off again when I'm asked to review
someone else's web site...]

At a quick count, about 60% of the CSS Zen layouts on the front page are
too small for me to read if I set the font-size to be how I like it at
100%.

Eg:
http://www.csszengarden.com/?cssfile=/176/176.css&page=0
http://www.csszengarden.com/?cssfile=/178/178.css&page=0

Now, these are not pages styled by amateurs.

That's true...

Their authors should be the ones who understand accessibility and
respecting user preferences more than anyone.

Oh, really? I'm sorry, but I have to disagree with this. The
csszengarden is a spectacular demonstration of what -can- be achieved with
CSS - but under what are practically laboratory conditions: it is in no
way a role model to be used, as it stands, for production use on the web.

Its contributors, like all of us, have their strengths and weaknesses, and
furthermore they are somewhat limited by the rules of the zengarden.

Given that very few of us are experts in all of graphic design,
typography, accessibility and web technologies, I'd risk saying that the
best web pages, by and large, are the product of co-operative work
between, at least, two people: a graphic designer with some understanding
of the web, and a techie with some sympathy for graphic design.

Another good point. Good at making things look good with CSS doesn't
necessarily mean they are using it correctly.

No.

 

[Alan J. Flavell]

If you know of a desktop TFT, 19 inches or bigger that does very high
pixels per inch please let me know. It's about time someone put a laptop
resolution panel a desktop casing.

Well, a web search suggests, if money is no object: ViewSonic VP2290b

A 22-inch display with resolution up to 3,840 x 2,400

And a price tag not far short of GBP 5K

The comparable IBM T221 display was stated to have been reduced from $13K
to $8K in 2002, but I can't find what it would cost nowadays (if still
available).


But seriously...

At any rate, there are TFT/LCD displays on the consumer market with dot
pitches from 0.297mm to 0.20mm, which in round numbers is a linear factor
of 1.5:1

That kind of factor sure makes a difference in the text display stakes,
just as it did in the CRT era when I compared an acquaintance's CRT
display set below 90dpi, against mine set at around 135dpi (to the disdain
of the peanut gallery who told me I had no business to set it higher than
the CRT's inherent dot pitch, but it worked for me).


Those high-res displays I mentioned at the start still cost an arm and a
leg, for sure. But I remember when, for example, a twelve-port 10Mbit
ethernet hub cost GBP1K: things change, and I suspect the video games
industry is a powerful driver for high-resolution fast-response displays
at an affordable price.

 

[Neredbojias]

With neither quill nor qualm, Jim Higson quothed:

Can't I do both?

Of course. In fact, the problem won't seem as serious if you do.

Ok, I'd say I'm capable at Javascript.
Check out my pet project beta (in Gecko)
http://81.5.150.113/wysi

?? I got this:

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head><title>Wikiwyg</title>
<link rel="stylesheet" type="text/css" href="skins/monobook/main.cssz"/>
<link rel="stylesheet" type="text/css" href="wysimain.cssz"/>
<script type="text/javascript" src="all.jamz"/><script
type="text/javascript">window.onload = init;</script>
<link rel="SHORTCUT ICON" href="favicon.ico"/>

It'll be released proper pretty soon.


This is interesting, but I'm not certain what they're doing is what I want.
They're resizing the text to always only just fill the page without
scrolling. I don't think this'd work very well for my 700 word articles.

No, it was meant as an example only. Put some text in a container,
gauge the height of said container with javascript, and resize the text
accordingly. The "example" proves the technique works.

 

[Toby Inkster]

<script type="text/javascript" src="all.jamz"/><script
type="text/javascript">window.onload = init;</script>

Why not add the line "window.onload=init;" to the end of all.jamz?

 

[Jim Higson]

With neither quill nor qualm, Jim Higson quothed:


Of course. In fact, the problem won't seem as serious if you do.


?? I got this:

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head><title>Wikiwyg</title>
<link rel="stylesheet" type="text/css" href="skins/monobook/main.cssz"/>
<link rel="stylesheet" type="text/css" href="wysimain.cssz"/>
<script type="text/javascript" src="all.jamz"/><script
type="text/javascript">window.onload = init;</script>
<link rel="SHORTCUT ICON" href="favicon.ico"/>
</head><body/></html>

Yeah, it's an experiment in fitting Ajax into MediaWiki. That's all you'll
get without Javascript. With Javascript you should get the full wiki, but
it's only been tested in recent versions of Gecko.

 

[Neredbojias]

With neither quill nor qualm, Toby Inkster quothed:

Why not add the line "window.onload=init;" to the end of all.jamz?

I tried but for some reason was unable to download all.jamz so...

-Er, wait a minute. You are replying to a line by the OP. Yes, I
agree, that would be logical.

 

[Spartanicus]

That kind of factor sure makes a difference in the text display stakes,
just as it did in the CRT era when I compared an acquaintance's CRT
display set below 90dpi, against mine set at around 135dpi (to the disdain
of the peanut gallery who told me I had no business to set it higher than
the CRT's inherent dot pitch, but it worked for me).

The name called reference would be me.

Contrary to what is alleged I did not tell Mr Flavell that "he had no
business to set it higher than the inherent dot pitch". What I did tell
Mr Flavell was that it's wrong to blame others for issues caused by the
misconfiguration of his own hardware.

Many people are confused about the issue of resolution on CRT monitors,
they think that by increasing the Screen Area setting, the resolution
follows suit. For colour CRT monitors this is only true up to a certain
point. Colour CRTs use a mask and phosphor clusters known as "dots",
these dots have a physical dimension referred to in specifications as
"dot pitch" [1]. A few examples: 0.24mm dot pitch = ~105PPI max
resolution, 0.27mm dot pitch = ~94PPI max resolution.

Mr Flavell has his Screen Area setting at a value considerably beyond
what his CRT is capable of displaying, the consequence of doing that is
that definition is lost. This makes text difficult to read, to
compensate he has configured his browser's default font size to a higher
setting. People who misconfigure their hardware like that then often
blame web authors for their problems.

Another issue that Mr Flavell is confused about are high resolution
screens. The 200PPI monitors referred to by him are not intended, or
suitable for general usage. Bitmapped resources (icons, buttons and
other UI elements, images etc.) shrink down to a size that is very
difficult to use. What Mr Flavell doesn't get is that these screens are
intended to be used in a dual monitor setup in conjunction with a
monitor with a normal resolution which displays the OS and the
application UI.

Such a setup is valuable for people who create work for print on
computer screens such as print cartographers. By using a 200PPI screen
to display the work they get a better view of what the result will be on
paper.

[1] More often than not the dot pitch figure listed by manufacturers is
not realistic, it may refer to the size of the grid pitch (the grid lies
a certain distance behind the phosphor dots), or it refers to the
vertical dot pitch only.

 

[Jim Higson]

Why not add the line "window.onload=init;" to the end of all.jamz?

Yep, that'd be better.
*shakes head at own bufoonery*

Incidently, all.jamz is gzipped output from jsjam, I get about 20:1
compression overall.

 

[Jim Higson]

With neither quill nor qualm, Toby Inkster quothed:


I tried but for some reason was unable to download all.jamz so...

This is odd. I just tried here and it was fine. The file is javascript with
content-type gzip, but apart from being quite long is nothing all that
unusual.

Can you give any more information, for example the browser used, time etc or
even a capture of the HTTP get request.

-Er, wait a minute. You are replying to a line by the OP. Yes, I
agree, that would be logical.

I also agree and I'm the OP!

 

[Jim Higson]

That kind of factor sure makes a difference in the text display stakes,
just as it did in the CRT era when I compared an acquaintance's CRT
display set below 90dpi, against mine set at around 135dpi (to the disdain
of the peanut gallery who told me I had no business to set it higher than
the CRT's inherent dot pitch, but it worked for me).

The name called reference would be me.

Contrary to what is alleged I did not tell Mr Flavell that "he had no
business to set it higher than the inherent dot pitch". What I did tell
Mr Flavell was that it's wrong to blame others for issues caused by the
misconfiguration of his own hardware.

Many people are confused about the issue of resolution on CRT monitors,
they think that by increasing the Screen Area setting, the resolution
follows suit. For colour CRT monitors this is only true up to a certain
point. Colour CRTs use a mask and phosphor clusters known as "dots",
these dots have a physical dimension referred to in specifications as
"dot pitch" [1]. A few examples: 0.24mm dot pitch = ~105PPI max
resolution, 0.27mm dot pitch = ~94PPI max resolution.

As an aside, I used to actually quite like the effect of a CRT at
resolutions above what the dot pitch said it could handle. So long as it
didn't bring down the refresh rate, the low-GPU fake antialiasing in games
looked fine, just like my 640*480 console looks less pixelated on a 23 inch
TV than my 1280x1024 computer does on a 19 inch monitor.

Mr Flavell has his Screen Area setting at a value considerably beyond
what his CRT is capable of displaying, the consequence of doing that is
that definition is lost. This makes text difficult to read, to
compensate he has configured his browser's default font size to a higher
setting. People who misconfigure their hardware like that then often
blame web authors for their problems.

Another issue that Mr Flavell is confused about are high resolution
screens. The 200PPI monitors referred to by him are not intended, or
suitable for general usage. Bitmapped resources (icons, buttons and
other UI elements, images etc.) shrink down to a size that is very
difficult to use.

Hmmm... what is the current state of vector interfaces? I think I read
somewhere GNOME uses lazy rasterised SVG for all icons etc. I might have
been reading about a future version, mind.

Maybe future very high res screens are where Opera will gain greater use,
because it can zoom the bitmapped parts as well as increase the text size.

What Mr Flavell doesn't get is that these screens are
intended to be used in a dual monitor setup in conjunction with a
monitor with a normal resolution which displays the OS and the
application UI.

Such a setup is valuable for people who create work for print on
computer screens such as print cartographers. By using a 200PPI screen
to display the work they get a better view of what the result will be on
paper.

[1] More often than not the dot pitch figure listed by manufacturers is
not realistic, it may refer to the size of the grid pitch (the grid lies
a certain distance behind the phosphor dots), or it refers to the
vertical dot pitch only.

 

[Neredbojias]

With neither quill nor qualm, Jim Higson quothed:

This is odd. I just tried here and it was fine. The file is javascript with
content-type gzip, but apart from being quite long is nothing all that
unusual.

Can you give any more information, for example the browser used, time etc or
even a capture of the HTTP get request.

I tried yesterday a.m. with both Moz 1.7.11 and Ie 6 on a windows xp
machine. Have no idea of the header as I still have a some confusion
with the term due to its other meaning.

I also agree and I'm the OP!

Well I'm the responding part, or RP, so Toby must be the pejorative
party, or...

 

[Alan J. Flavell]

Spartanicus, who had requested me to killfile him, had been scribbling

behind my back, and making various assumptions which go way beyond the
testimony that is in evidence. But I shall make allowances, and respond
to a few points despite that.

[...]
I was responding to a question about the availability of high resolution
screens. Is that so upsetting?

At the price, it's hardly surprising.

Sure. Which is why Microsoft have an article explaining their OS features
which *compensate* for that.

It's one way of dealing with the situation, certainly. Both Microsoft and
IBM, in their various offerings, demonstrate that it's not the only way.

Anyway: how difficult was it to discern that my discussion of these beasts
in relation to the WWW, given their price, was whimsical at the moment?
This line below my discussion:

|| But seriously...

might have given some kind of clue.

But I've no doubt that things *will* develop in that direction, and, as
long as some authors are fixated on designs that only work at about
100dpi, there's going to come a point where the flexibility snaps.

Meantime, perhaps Spartanicus would care to write to Iiyama and lecture
them in corresponding terms about their CRT displays, since, according to
them, their Visionmaster Pro 454 19" display (e.g HM903DTB) supports
resolutions (their term) up to 2048x1536. (Corresp. display area given as
366x274mm, for those who are calculating in the margin). Not that it
matters, but I wasn't going that far with mine.

bon soir