An IE font usage problem...

I've got a problem that IE doesn't seem to use the entire font unicode
subgroups from a hinted font file. When I display the following web page in
IE 6 as opposed to Firefox 1.0 PR.

Check out the web page in question...

http://clientserver.home.comcast.net/unicode.html

If you examine, for a few examples, characters such as square root
(&#8730, prime (&#8242 and double prime (&#8243.

At the bottom of the page, in the legend, I've included the hints for
columns A, B, and C...

I also see the same problems, with hinted print fonts, when I do a print
preview of the web page.

So, any ideas why IE is puking on these?

Firefox seems to do a beautiful job of using all the available characters
and then seems to provide, were some characters are missing, excellent
substitutions from other installed fonts.

Rich
--
RichardRPlourde (AT) comcast (DOT) net
http://RichardRPlourde.home.comcast.net

"Richard R Plourde" <SeaPlusPlus (AT) hotmail (DOT) com> wrote

Check out the web page in question...

http://clientserver.home.comcast.net/unicode.html

In addition to the previous post...

Using Windows/XP Character Map program it seems to be the Unicode subrange
"Mathematical Operators" which appears to be missing the characters...

Don't know if that's a hint to what I'm doing wrong.

--
Rich
RichardRPlourde (AT) comcast (DOT) net
http://RichardRPlourde.home.comcast.net

On Sun, 31 Oct 2004, Richard R Plourde wrote:

If you examine, for a few examples, characters such as square root
(&amp;#8730, prime (&amp;#8242 and double prime (&amp;#8243.

So, any ideas why IE is puking on these?

One of two reasons:

1) the font doesn't claim to support the character range in question,
but your browser configuration hasn't configured an appropriate font
for fallback.

2) the font -does- claim to support the character range in question,
but its support is incomplete. In this situation, IE seemingly makes
no effort to hunt the missing characters.

Simple recommendation: don't specify font faces for obscure
characters. Users -may- need to tune their browser configuration
(especially IE).

When you understand IE's behaviour better than I do, you /might/ get
marginally better results with a very carefully chosen font
specification (see discussions between myself and A.Prilop in these
usenet groups). But I'm still sceptical.

The actual -repertoire- of various fonts distributed with MS OSes
varies from one OS version to another, even without change in the name
of the font. So don't assume that the repertoire that you tested on
XP (for example) will also work for '2000, and certainly not for NT4.
But Lucida Sans Unicode, or Arial Unicode MS if the user has it, are
good choices for difficult characters.

Do -not- on any account put the generic sans-serif or serif at the end
of your CSS font list, in such a situation.

Firefox seems to do a beautiful job of using all the available characters

No surprises there.

On Sun, 31 Oct 2004, Richard R Plourde wrote:

Don't know if that's a hint to what I'm doing wrong.

You're not doing anything wrong, in principle. It's IE's shortcomings
that you need to workaround.

Sorry, I omitted to include my relevant URL in the previous posting,
so here it is:

http://ppewww.ph.gla.ac.uk/~flavell/...ers-fonts.html

On Sun, 31 Oct 2004, Richard R Plourde wrote:

If you examine, for a few examples, characters such as square root
(&amp;#8730, prime (&amp;#8242 and double prime (&amp;#8243.

So, any ideas why IE is puking on these?

"Alan J. Flavell" <flavell (AT) ph (DOT) gla.ac.uk> wrote

One of two reasons:

1) the font doesn't claim to support the character range in question,
but your browser configuration hasn't configured an appropriate font
for fallback.

The font... Georgia does in fact have the characters... Firefox uses them
(IE doesn't).

2) the font -does- claim to support the character range in question,
but its support is incomplete. In this situation, IE seemingly makes
no effort to hunt the missing characters.

The characters are not missing... Character Map shows them... they are in
the font.

Simple recommendation: don't specify font faces for obscure
characters. Users -may- need to tune their browser configuration
(especially IE).

For what I want to do these are NOT 'obscure characters', so I'll just
include on my web site strong recommendations for all recent browsers other
than IE. ;-)

At least IE chooses 'appropriate' (funky looking) characters... IE doesn't
use the characters which provide professional loooking typography.

When you understand IE's behaviour better than I do, you /might/ get
marginally better results with a very carefully chosen font
specification (see discussions between myself and A.Prilop in these
usenet groups). But I'm still sceptical.

The actual -repertoire- of various fonts distributed with MS OSes
varies from one OS version to another, even without change in the name
of the font. So don't assume that the repertoire that you tested on
XP (for example) will also work for '2000, and certainly not for NT4.
But Lucida Sans Unicode, or Arial Unicode MS if the user has it, are
good choices for difficult characters.

I will try those two unicode fonts and subsequently may hint these but I'm
hinting three separate display and three sepatate print fonts.

For my purposes 'Courier New' is used in both display and print.

Do -not- on any account put the generic sans-serif or serif at the end
of your CSS font list, in such a situation.

This I will do also. I thought it was imperative that they should be
included. If they aren't there, people say, it's an error not to include
them.

Firefox seems to do a beautiful job of using all the available
characters

No surprises there.

Ain't life grand... ;-)

Maybe Longhorn will totally turn its back on IE ! ! ! ! ;-)

Thanks...

Rich

On Sun, 31 Oct 2004, Richard R Plourde wrote:

1) the font doesn't claim to support the character range in
question, but your browser configuration hasn't configured an
appropriate font for fallback.

The font... Georgia does in fact have the characters...

But that's not the point! The font does not claim to support them, so
IE will look for some better-populated font.

Font properties extension -> charset/unicode reports, for this version
of Georgia at least (that's Win/2000 Pro, font version 2.05) that it
supports only:

Basic Latin
Latin-1 Supplement
Latin Extended-A
Greek
Cyrillic

Whereas if we take say Lucida Sans Unicode, it has lots of extra
ranges supported, including IPA Extensions, Combining Diacritical
Marks, ect. etc. and in particular: Mathematical Operators.

Firefox uses them (IE doesn't).

I don't know the internals of IE in detail: but the observations fit
my explanation well, I'd say.

2) the font -does- claim to support the character range in question,
but its support is incomplete. In this situation, IE seemingly makes
no effort to hunt the missing characters.

The characters are not missing...

In this case it *was* point (1) which applied, so point (2) is
irrelevant. In other situations, you'd find the other point coming
into play.

Character Map shows them...

Doesn't help. Font properties extension -> Charset/Unicode tells a
different story, so IE won't use it.

they are in the font.

Doesn't help. IE won't use it.

Simple recommendation: don't specify font faces for obscure
characters. Users -may- need to tune their browser configuration
(especially IE).

For what I want to do these are NOT 'obscure characters',

Well, feel free to supply your own descriptive phrase, it makes no
difference to the end result: IE has problems rendering these
characters with the fonts that you want it to use. That's the bottom
line.

so I'll just include on my web site strong recommendations for all
recent browsers other than IE. ;-)

You could tell IE users that they'll get better results (as far as the
repertoire is concerned - the cosmetics will no doubt leave something
to be desired) by checking the "Ignore font styles specified on web
pages" box on the tools->accessibility, and configuring their browser
to default to a well-populated font.

For my purposes 'Courier New' is used in both display and print.

Monospaced fonts proved to be an additional problem, as my web page
explains. MS's monospaced fonts (as delivered with the OS) have an
even worse character repertoire coverage.

Do -not- on any account put the generic sans-serif or serif at the end
of your CSS font list, in such a situation.

This I will do also. I thought it was imperative that they should be
included.

Oh no; it's often recommended, for typographical reasons, but the two
criteria are pulling in opposite directions in this instance: you can
either have better typography (with missing glyphs) or better
character repertoire (with somewhat klunky typography).

If they aren't there, people say, it's an error not to include
them.

Well, not exactly - it's a recommendation, but by no means compulsory,
and in this case it turns out to be harmful (to IE, anyway). Still,
as I say, IE users might get better results by disabling the
author-proposed fonts and making their own best choice.

good luck

On Sun, 31 Oct 2004, Richard R Plourde wrote:

I've got a problem that IE doesn't seem to use the entire font unicode
subgroups from a hinted font file.
[...]
At the bottom of the page, in the legend, I've included the hints for
[...]
I also see the same problems, with hinted print fonts,

You confuse me. Please read
http://partners.adobe.com/asn/tech/type/hinting.jsp
http://developer.apple.com/fonts/TTQ...QS03/FQS3.html
http://www.microsoft.com/typography/...tingIntro.mspx
to learn (more?) about font hinting.

--
Top-posting.
What's the most irritating thing on Usenet?

On Sun, 31 Oct 2004, Richard R Plourde wrote:

I've got a problem that IE doesn't seem to use the entire font unicode
subgroups from a hinted font file.
[...]
At the bottom of the page, in the legend, I've included the hints for
[...]
I also see the same problems, with hinted print fonts,

"Andreas Prilop" <nhtcapri (AT) rrzn-user (DOT) uni-hannover.de> wrote

You confuse me. Please read
http://partners.adobe.com/asn/tech/type/hinting.jsp
http://developer.apple.com/fonts/TTQ...QS03/FQS3.html
http://www.microsoft.com/typography/...tingIntro.mspx
to learn (more?) about font hinting.

No, Andreas, I'm sorry what I wrote is ambiguous... sorry...

When I said hinted I meant as suggested in my CSS font statements... i.e.

font: medium Georgia serif;

I remember reading that these suggestions are just hints to the browser and
not expected to be followed blindly by all browsers the same way.

Sorry if I confused the issue.

Rich
--
Thanks a lot...

Rich

RichardRPlourde (AT) comcast (DOT) net
http://RichardRPlourde.home.comcast.net

"Richard R Plourde" <SeaPlusPlus (AT) hotmail (DOT) com> wrote

On Sun, 31 Oct 2004, Richard R Plourde wrote:

I've got a problem that IE doesn't seem to use the entire font unicode
subgroups from a hinted font file.
[...]
At the bottom of the page, in the legend, I've included the hints for
[...]
I also see the same problems, with hinted print fonts,

"Andreas Prilop" <nhtcapri (AT) rrzn-user (DOT) uni-hannover.de> wrote in message
news:Pine.GSO.4.44.0411011602130.6794-100000 (AT) s5b004 (DOT) ..

You confuse me. Please read
http://partners.adobe.com/asn/tech/type/hinting.jsp
http://developer.apple.com/fonts/TTQ...QS03/FQS3.html
http://www.microsoft.com/typography/...tingIntro.mspx
to learn (more?) about font hinting.

No, Andreas, I'm sorry what I wrote is ambiguous... sorry...

When I said hinted I meant as suggested in my CSS font statements... i.e.

font: medium Georgia serif;

If you used this particular line, you'd need a comma between the alternate
font family names.

On Sun, 31 Oct 2004, Richard R Plourde wrote:

The font... Georgia does in fact have the characters...

"Alan J. Flavell" <flavell (AT) ph (DOT) gla.ac.uk> wrote

But that's not the point! The font does not claim to support them, so
IE will look for some better-populated font.

Font properties extension -> charset/unicode reports, for this version
of Georgia at least (that's Win/2000 Pro, font version 2.05) that it
supports only:

Basic Latin
Latin-1 Supplement
Latin Extended-A
Greek
Cyrillic

Whereas if we take say Lucida Sans Unicode, it has lots of extra
ranges supported, including IPA Extensions, Combining Diacritical
Marks, ect. etc. and in particular: Mathematical Operators.

Okay, I see exactly what you mean. I was wondering where the 'disconnect'
was and now I know.

Thanks.

I've (sort of) fixed the problem... actually a compromise...

I'm using the three fonts (serif, sans-serif, and monospace) thus...
______________________

SERIF

Prose (headings, lists, table data, term descriptions, and normal text)...

Display Font -> font: medium Georgia; /* serif */

Print Font -> font: 10.5pt normal "Times New Roman"; /* small pica serif */

Tradeoffs...

Georgia is the font that IE won't use the Mathematical Operators so use
sans-serif for them... actually I believe math looks better in sans-serif.
______________________

SANS-SERIF

Mathematical expressions (equations, variables, constants, and some source
code listings; also used for menu, table lables and description terms)...

Display Font -> font: medium "Lucida Sans Unicode"; /* sans-serif */

Print Font -> font: 10.5pt "Lucida Sans Unicode"; /* small pica sans-serif
*/


Tradeoffs...

Lucida Sans Unicode has a section of ugly characters around "greater than or
equal to" and "less than or equal to" characters and also there is no bold
or italics.
______________________

MONOSPACE

Columnar presentations (numerical tables, printouts, and some source code
listings -- wrap and no wrap)...

Display Font -> font: medium "Courier New"; /* monospace */

Print Font -> font: 10.5pt "Courier New"; /* small pica monospace */


Tradeoffs...

Courier New has no known tradeoffs.
______________________

You could tell IE users that they'll get better results (as far as the
repertoire is concerned - the cosmetics will no doubt leave something
to be desired) by checking the "Ignore font styles specified on web
pages" box on the tools->accessibility, and configuring their browser
to default to a well-populated font.

Hopefully the method I'm using now will make this not neccessary.

Monospaced fonts proved to be an additional problem, as my web page
explains. MS's monospaced fonts (as delivered with the OS) have an
even worse character repertoire coverage.

My use of monspace as descibed above should be very 'light'... and this
shouldn't be a problem.

Do -not- on any account put the generic sans-serif or serif at the end
of your CSS font list, in such a situation.

This I will do also. I thought it was imperative that they should be
included.

Oh no; it's often recommended, for typographical reasons, but the two
criteria are pulling in opposite directions in this instance: you can
either have better typography (with missing glyphs) or better
character repertoire (with somewhat klunky typography).

I think the method above is a good compromise. Now I need to see what the
typography looks like on as many systems as possible.

as I say, IE users might get better results by disabling the
author-proposed fonts and making their own best choice.

I'm hoping this won't be neccessary.

I'm now working on getting equations to be rendered reasonably well.

--
Thanks a lot...

Rich

RichardRPlourde (AT) comcast (DOT) net
http://RichardRPlourde.home.comcast.net