User selectable font size and IE6

[Paul Davis]

Hi all,
I'm building some style sheets and trying to play the old game of
balancing designer pixel perfection and still allowing users to adjust
their font sizes. The compromise I've made with the designer is that
he'll accept allowing users to change font sizes it it "looks right"
with the default setting.
I'm using em to set the font size. So, anyway, here's the problem. At
0.96em we get the perfect font size. But, if the font is changed with
'view->text size->larger' it becomes gigantic, 'view->text
size->smaller' makes it microscopic. Of course Firefox does nice
gradual changes with ctl+ and ctl-

Any thoughts on how to make this work better?

[Beauregard T. Shagnasty]

Paul Davis wrote:

Hi all,
I'm building some style sheets and trying to play the old game of
balancing designer pixel perfection and still allowing users to
adjust their font sizes. The compromise I've made with the designer
is that he'll accept allowing users to change font sizes it it "looks
right" with the default setting.

It sounds as if your designer doesn't care to allow visitors with less
than perfect eyesight to use your site.

I'm using em to set the font size. So, anyway, here's the problem. At
0.96em we get the perfect font size. But, if the font is changed with
'view->text size->larger' it becomes gigantic, 'view->text
size->smaller' makes it microscopic. Of course Firefox does nice
gradual changes with ctl+ and ctl-

Any thoughts on how to make this work better?

It has been written that IE has a problem/bug when font-sizing is done
using em units. Try: font-size: 96%; instead.

--
-bts
-Motorcycles defy gravity; cars just suck

[Jukka K. Korpela]

Scripsit Paul Davis:

I'm using em to set the font size.

Why? It sounds you are trying to set the _overall_ font size. Setting it in
em units (or as a percentage) can only have two effects:
1) none (user-selected font size is used)
2) font size is set to a value that systematically differs from what the
user has chosen, i.e. no matter what the user has done, the dee-ziner has
decided it's wrong.

So, anyway, here's the problem.

Yes. So stop creating it.

At 0.96em we get the perfect font size.

For some very odd value for "perfect". It is very unlikely that there is
_any_ effect, unless, of course, there's something you haven't told us.
Typical user choices like 12pt or 14pt will be used, since reducing the
declared size by 4% has no effect on the actual size - there's most probably
no font that is just a very little smaller than 12pt or 14pt.

But, if the font is changed with
'view->text size->larger' it becomes gigantic, 'view->text
size->smaller' makes it microscopic.

For some odd values for "gigantic" or "microscopi c". Granted, the IE font
control is primitive and coarse, but why, oh why, do so many authors want to
deprive users of even that primitive tool or to make it work in
unpredictable ways?

Any thoughts on how to make this work better?

Remove all font-size declarations (except for headings).

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

[Paul Davis]

Beauregard T. Shagnasty wrote:

Paul Davis wrote:

Hi all,
I'm building some style sheets and trying to play the old game of
balancing designer pixel perfection and still allowing users to
adjust their font sizes. The compromise I've made with the designer
is that he'll accept allowing users to change font sizes it it "looks
right" with the default setting.

It sounds as if your designer doesn't care to allow visitors with less
than perfect eyesight to use your site.

the trouble is its true. I haven't many who do.
You know its a sad day when the developer is fighting the designer to
make a site *more usable*

>

I'm using em to set the font size. So, anyway, here's the problem. At
0.96em we get the perfect font size. But, if the font is changed with
'view->text size->larger' it becomes gigantic, 'view->text
size->smaller' makes it microscopic. Of course Firefox does nice
gradual changes with ctl+ and ctl-

Any thoughts on how to make this work better?

It has been written that IE has a problem/bug when font-sizing is done
using em units. Try: font-size: 96%; instead.

That seems to work.
I Think I found a piece of the problem too.
The font size was being defined in the rule
* {
font-size: 0.96em;
font-family: Arial, Helvetica, sans-serif;
}
so, apparently the variation in font size gets magnified for each level
of tag nesting.
Doh !!!
Font size is now in the body definition, all is well.

>
--
-bts
-Motorcycles defy gravity; cars just suck

[Beauregard T. Shagnasty]

Paul Davis wrote:

Beauregard T. Shagnasty wrote:

>It sounds as if your designer doesn't care to allow visitors with
less than perfect eyesight to use your site.

the trouble is its true. I haven't many who do. You know its a sad day
when the developer is fighting the designer to make a site *more
usable*

There are many sad days. Most "deeziners" don't have a clue about
accessibility and usability. Every deziner I know has perfect eyesight.

>It has been written that IE has a problem/bug when font-sizing is done
using em units. Try: font-size: 96%; instead.

That seems to work.
I Think I found a piece of the problem too.
The font size was being defined in the rule
* {
font-size: 0.96em;
font-family: Arial, Helvetica, sans-serif;
}
so, apparently the variation in font size gets magnified for each level
of tag nesting.
Doh !!!
Font size is now in the body definition, all is well.

Ah yes. Glad you found the problem.

--
-bts
-Motorcycles defy gravity; cars just suck

[Paul Davis]

Jukka K. Korpela wrote:

Scripsit Paul Davis:

I'm using em to set the font size.

Why? It sounds you are trying to set the _overall_ font size. Setting it in
em units (or as a percentage) can only have two effects:
1) none (user-selected font size is used)
2) font size is set to a value that systematically differs from what the
user has chosen, i.e. no matter what the user has done, the dee-ziner has
decided it's wrong.

Re-read the post, the goal is to ensure users can change it.
It doesn't matter if the designer is wrong. The business owners decided
to go with the design, they decided to pay me to implement.

>

So, anyway, here's the problem.

Yes. So stop creating it.

At 0.96em we get the perfect font size.

For some very odd value for "perfect". It is very unlikely that there is
_any_ effect, unless, of course, there's something you haven't told us.
Typical user choices like 12pt or 14pt will be used, since reducing the
declared size by 4% has no effect on the actual size - there's most probably
no font that is just a very little smaller than 12pt or 14pt.

But, if the font is changed with
'view->text size->larger' it becomes gigantic, 'view->text
size->smaller' makes it microscopic.

For some odd values for "gigantic" or "microscopi c". Granted, the IE font
control is primitive and coarse, but why, oh why, do so many authors want to
deprive users of even that primitive tool or to make it work in
unpredictable ways?

Any thoughts on how to make this work better?

Remove all font-size declarations (except for headings).

Great idea, with a small problem.
Not being independently wealthy, I need to earn a paycheck to feed my
family. My job as a developer is to take a designer's idea and turn it
into reality. If I don't do that, management will hire someone who
will.
I am not a designer, I don't claim to be and I'm not artistic enough to
do it even if I wanted to. I write code.
While your suggestions are nice from a purity standpoint (although it
suggests that we should all use gopher browsers) It doesn't help from
an employment standpoint.

>
--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

[Michael Winter]

Paul Davis wrote:

Beauregard T. Shagnasty wrote:

[snip]

>It has been written that IE has a problem/bug when font-sizing is done
using em units. Try: font-size: 96%; instead.

[snip]

The font size was being defined in the rule
* {
font-size: 0.96em;
font-family: Arial, Helvetica, sans-serif;
}
so, apparently the variation in font size gets magnified for each level
of tag nesting.

The effect of that rule is that the computed font size for each element
does gradually get smaller the deeper the element is within the document
tree, but that isn't why MSIE fouls up text resizing so badly.

When em length font sizes are applied to an element and a descendant
without a non-em length font-size declaration applied to an ancestor,
MSIE grossly miscalculates the computed font size for the descendant.

I demonstrated this to someone else earlier this year (the text was from
another example):

<http://mwinter.webhop. info/dd/john-scheuer/font-size.html>

If a non-em length font-size declaration was included in the rule for
the body element or the html element, or if the 1em value in the
#em-length rule was changed to 100%, the problem would disappear.

Mike

[Jukka K. Korpela]

Scripsit Paul Davis:

Re-read the post, the goal is to ensure users can change it.

You cannot ensure it and you cannot prevent it. You can just make it more
difficult, and that seems to be what you have originally decided to do.

Regarding reading, you seem to have quoted almost my entire message, for no
good reason since you don't comment all of it, and you haven't apparently
read it comprehensively .

It doesn't matter if the designer is wrong.

It matters to me, and to the user, and ultimately to the site owner too.

The business owners
decided to go with the design, they decided to pay me to implement.

Fine. Are you suggesting that you made an agreement on doing something that
shouldn't be done _and_ you don't have an idea of how to do that? Do you
expect us to help you with that for free? Anyway, it's pointless since we
don't even know what you really agreed to do, so we can't decide whether
it's even possible or not.

My job as a developer is to take a designer's idea and turn it
into reality.

Even if you haven't even understood the idea and don't know how to implement
it? If someone pays you for that, shouldn't you get a subcontractor now?
(Please don't contact me for that. You would most probably have to pay me
much more than you get.)

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

[Andy Dingley]

Paul Davis wrote:

I'm building some style sheets and trying to play the old game of
balancing designer pixel perfection and still allowing users to adjust

I suggest balancing the dezyner outside a high window.

I'm using em to set the font size. So, anyway, here's the problem. At
0.96em we get the perfect font size

At 1em your users get their perfect font size.

Any thoughts on how to make this work better?

I used to think breaking the dezyner's fingers was the way to go, but
now I favour poking them in the eyes -- just enough to reduce their
vision until they appreciate the problem.
BODY TEXT GOES AT 1EM

NOTHING GOES LESS THAN 60% / 0.6em

If you want the text to fit the hole, size the hole in ems too

Set your heights in ems. Probably the widths too. Keep your widths in
pixels if you're having to work around lots of fixed size images. The
text in ems will wrap to fit better than the images will fit a box
sized in ems. Avoid mixing units in adjacent boxes though -- stick to
one or the other (or at least test carefully). Mixing units for
horizontal and vertical is OK. Mixing units between parents and
children is OK.

If you absolutely _must_ keep some font sizes fixed (small breadcrumbs
and flyspeck-3 are reasonable) then size the text in pixels, size the
container height in pixels, and hope they don't view it in Firefox.

[Bergamot]

Jukka K. Korpela wrote:

Scripsit Paul Davis:

>But, if the font is changed with
'view->text size->larger' it becomes gigantic, 'view->text
size->smaller' makes it microscopic.

For some odd values for "gigantic" or "microscopi c".

Not 'odd' at all, at least not the way you usually infer that means.
Windows IE has a well-known bug scaling font sizes set in em units. Any
browser View->Text Size setting other than "Medium" truly does end up
gigantic or microscopic. That bug has been mentioned in this and other
newsgroups many times. I'm surprised you aren't aware of it.

--
Berg