Fonts

[Beauregard T. Shagnasty]

Using a WeirdSoftware filter?

<lol>

I just found the reason. !delete Header 64.59.135.176

...and Gufus' GG header: NNTP-Posting-Host: 64.59.135.176

That's definitely a Shaw Cable IP address. Somewhere in the past,
someone from Shaw must have been flooding Usenet with drivel, and I
binned him.

 

[Blinky the Shark]

<lol>

I just found the reason. !delete Header 64.59.135.176

..and Gufus' GG header: NNTP-Posting-Host: 64.59.135.176

That's definitely a Shaw Cable IP address. Somewhere in the past,
someone from Shaw must have been flooding Usenet with drivel, and
I binned him.

Nice to see someone else that knows "drivel". Like those with
problems with there/they're/their and your/you're, it seems like
most people using Usenet think it's "dribble".

 

[Beauregard T. Shagnasty]

Nice to see someone else that knows "drivel". Like those with
problems with there/they're/their and your/you're, it seems like most
people using Usenet think it's "dribble".

Sometimes I dribble, usually when reading drivel.

 

[Jukka K. Korpela]

Scripsit Bergamot:

Or here's a novel idea: leave the family out altogether and let your
visitors decide for themselves what font to use.

In special cases, that idea might fire back. Suppose that you use characters
that do not commonly appear in fonts that are used as default fonts by
browsers. Say, the diameter sign (U+2300, occurring fairly often in
technical texts but seldom in fonts) or "h" with caron (U+021F, probably
used only in Romani (= gypsee) language as written in Finland) or the ezh
letter (U+0292, used as a phonetic symbol in IPA and as a letter in some
languages). If you don't set the font family at all in your CSS code (or in
HTML markup), _most_ users will just symbols of unrepresentable characters.
That's because they are using IE with Times New Roman as the default font.

On Firefox, the users will probably see all the characters, but many of them
will appear as picked up from fonts other than the rest. Such things happen
even in printed matter, for quite similar reasons. They can be a typographic
nightmare.

Surely they can fix the situation by changing the browser's default font. Do
they know how to do that? Will they do that? If they switch the font to,
say, Arial Unicode MS (assuming they have it), they will probably see all
the characters. But then they will have to switch it back to something else,
since otherwise _all_ pages that do not specify font family (and that means
a lot of pages that concentrate on presenting useful content) will appear in
that font, which is fairly dull - and has no real italics or real bold face.
(Does this matter? It does. Fake italics look real bad and even wrong for
many characters, e.g. "\" or Cyrillic letters.)

If, on the other hand, the author checks the character repertoire he uses
and writes down a font-family list containing fonts so that each of them has
all the characters needed, such as
body { font-family: Code2000, Arial Unicode MS, Lucida Sans Unicode }
then most users will see all the characters properly. Perhaps not in the
best possible font, but you can't win them all.

 

[Jukka K. Korpela]

Scripsit Animesh K:

Your wording is odd, since you confuse fonts with characters or glyphs and
diacritics with letters containing them, but you are basically right:
Trebuchet MS lacks glyphs for some of the characters used on the page, such
as n with dot above. Therefore, browsers either display a symbol denoting an
unrepresentable character or pick up a character from another font. (A
clever browser could construct a representation, by decomposing a character
into a base character and a combining diacritic mark. This would be quite in
accordance with the Unicode standard, but it's probably high above the heads
of browser designers, even though it would be almost trivial to implement.)

And here the same text displays fine in IE 6.0 with Tahoma

http://www.stutimandal.com/new/poemgen.php?id=33

Maybe it does, maybe it does not. On my computer, Tahoma contains n with dot
above, so I see the text in one font. On someone else's computer, maybe not.
Maybe their computer lacks Tahoma, or maybe their Tahoma is different. At
least the page
http://www.fileformat.info/info/unicode/char/1e45/fontsupport.htm
does not list Tahoma as one of the fonts supporting that character.

The secondary font suggested on the page, Times, is among the worst possible
choices. The Times font, where available (it's _not_ the same as Times New
Roman!) tends to contain a very small set of characters. On my computer, for
example, Times is a Type 1 font supporting little more than the Windows
Latin 1 repertoire - e.g., even Latin 2 characters are missing.

Moreover, Tahoma is basically condensed Verdana, suffering from the same
problems as Verdana _and_ from the condensation problem (though in some
situations, for some people, being condensed might be an asset). That is,
Tahoma has the same basic shapes of characters but horizontally denser. You
might try to approach this issue by using letter-spacing: 1px to make Tahoma
less dense.

So something like
font-family: Arial Unicode MS, Tahoma, Code2000, Everson Mono Unicode
would be better. (The letter-spacing idea is not good here, since CSS does
not let you say "if Tahoma is used, then...", so any letter-spacing set
would apply no matter which font is used.)

 

[mrcakey]

Actually IE 6.0 has has problems in displaying diacritics (characters with
a dot-below the alphabet) with "Trebuchet MS". I prefer Tahoma over
Trebuchet for that reason.

I would love to use Trebuchet MS on my webpage, if Uncle Gates insists IE
6.0 to support the diacritics.

Internet Explorer. Problem? I shall write to Bill Gates immediately. He
must want to know!

+mrcakey

 

[Animesh K]

In my Safari 2 and FF (Mac):

"posted by Stutiman?d?ala @ 12:17 PM"

except that the ? is a small bordered unfilled square.

Mostly fine though. Mac IE (perhaps not too surprisingly) is not
up to the job at all.

Err I am aware of this issue on the Mac. It looks embarassing, yes. But,
I haven't looked into fixing it yet since: 1) I don't have a Mac
(neither do I wish to overpay for one), 2) My site-traffic comes mostly
from Windows/Linux computers.

I think I need to fiddle with fonts till I get it right (which is
somewhat hard at Blogger, since the CSS file isn't completely in my hand
-- I need to look for a method to hack it).

Thanks though,
Animesh

 

[Animesh K]

Scripsit Animesh K:


Your wording is odd, since you confuse fonts with characters or glyphs
and diacritics with letters containing them, but you are basically
right: Trebuchet MS lacks glyphs for some of the characters used on the
page, such as n with dot above. Therefore, browsers either display a
symbol denoting an unrepresentable character or pick up a character from
another font. (A clever browser could construct a representation, by
decomposing a character into a base character and a combining diacritic
mark. This would be quite in accordance with the Unicode standard, but
it's probably high above the heads of browser designers, even though it
would be almost trivial to implement.)

You made my thoughts precise. I am not a font-expert, so I don't know
the various terms related to it.

FF on windows is a clever browser by your definition. Same is true for
Opera.

I wonder why FF on Mac screws it up (as Dorayme reported).

Maybe it does, maybe it does not. On my computer, Tahoma contains n with
dot above, so I see the text in one font. On someone else's computer,
maybe not. Maybe their computer lacks Tahoma, or maybe their Tahoma is
different. At least the page
http://www.fileformat.info/info/unicode/char/1e45/fontsupport.htm
does not list Tahoma as one of the fonts supporting that character.

The secondary font suggested on the page, Times, is among the worst
possible choices. The Times font, where available (it's _not_ the same
as Times New Roman!) tends to contain a very small set of characters. On
my computer, for example, Times is a Type 1 font supporting little more
than the Windows Latin 1 repertoire - e.g., even Latin 2 characters are
missing.

Moreover, Tahoma is basically condensed Verdana, suffering from the same
problems as Verdana _and_ from the condensation problem (though in some
situations, for some people, being condensed might be an asset). That
is, Tahoma has the same basic shapes of characters but horizontally
denser. You might try to approach this issue by using letter-spacing:
1px to make Tahoma less dense.

So something like
font-family: Arial Unicode MS, Tahoma, Code2000, Everson Mono Unicode
would be better. (The letter-spacing idea is not good here, since CSS
does not let you say "if Tahoma is used, then...", so any letter-spacing
set would apply no matter which font is used.)

May be I should switch to Arial. To me, Tahoma is a nice read and it
resembles the sans-serif font of Latex closest. That's why I chose it in
the first place.

I will experiment more with these inputs, though. Thanks

Animesh

 

[Animesh K]

In my Safari 2 and FF (Mac):

"posted by Stutiman?d?ala @ 12:17 PM"

except that the ? is a small bordered unfilled square.

Mostly fine though. Mac IE (perhaps not too surprisingly) is not
up to the job at all.

How does the page at Stutimandal renders? Do you see boxes here too?

http://www.stutimandal.com/new/poemgen.php?id=33

I did use a screenshot generating service, and that seemed to work on
many browsers.

Best regards,
Animesh

 

[Jukka K. Korpela]

Scripsit Animesh K:

FF on windows is a clever browser by your definition. Same is true for
Opera.

No, my "definition" said: "A clever browser could construct a
representation, by decomposing a character into a base character and a
combining diacritic mark." I don't think any browser does that. It would
mean that when a browser sees, say, n with dot above as a single precomposed
character and finds out that it hasn't got a glyph for it, it decomposes it
into plain n (which it has got) and combining dot above (which it might
have).

What FF and Opera do is that they pick up the precomposed character from
another font. Just what IE 7 does in this case, but probably using a
different replacement font. (Sorry, I'm too tired to check now.)

May be I should switch to Arial.

Hardly. Arial is much more limited than Arial Unicode MS.

To me, Tahoma is a nice read and it
resembles the sans-serif font of Latex closest.

I have almost learned to hate Tahoma. Especially in 12 pt size, it looks far
too dense, and in larger sizes it is, well, too large for most purposes
(excluding headings).

 

[dorayme]

Err I am aware of this issue on the Mac.

So this particular blemish does not appear so on FF on Windows?

 

[dorayme]

How does the page at Stutimandal renders? Do you see boxes here too?

http://www.stutimandal.com/new/poemgen.php?id=33

This looks fine, no boxes. Perhaps I have not the font on my own
machine that is showing those boxes I referred to in regard to

http://stutimandal.blogspot.com/2007/09/totakastakam.html

You have a few css sheets and I have not studied which font you
are instructing the copyright line to render in?

 

[Animesh K]

So this particular blemish does not appear so on FF on Windows?

Yes this blemish does not appears in FF.

My pages are first tested in FF and Opera on Windows. Then I try as much
as I can to fix any display issues with IE 6.0. I am planning to leave
that browser in future design (even though some 60% traffic comes using
IE 6.0 --- that's the only way to change some lazy bones to firefox .. I
wish more people would do that).

Thanks for the help.

 

[Animesh K]

This looks fine, no boxes. Perhaps I have not the font on my own
machine that is showing those boxes I referred to in regard to

http://stutimandal.blogspot.com/2007/09/totakastakam.html

You have a few css sheets and I have not studied which font you
are instructing the copyright line to render in?

The blogspot interface allows me to add a post with fonts. I started
with Trebuchet (hey it looks good!) and then after 3 months I checked to
see an error with IE 6.0. I cannot go back and fix 100 posts just
because IE 6.0 refuses to co-operate.

I didn't put those CSS sheets. They are developed for blogger.com's website.

I am unsure which font is being used for copyright line :-S

Well as long as my website (where I can control things) works better!

 

[Animesh K]

Scripsit Animesh K:


No, my "definition" said: "A clever browser could construct a
representation, by decomposing a character into a base character and a
combining diacritic mark." I don't think any browser does that. It would
mean that when a browser sees, say, n with dot above as a single
precomposed character and finds out that it hasn't got a glyph for it,
it decomposes it into plain n (which it has got) and combining dot above
(which it might have).

What FF and Opera do is that they pick up the precomposed character from
another font. Just what IE 7 does in this case, but probably using a
different replacement font. (Sorry, I'm too tired to check now.)

You are correct. FF and Opera are substituting the font by some other
font with no noticeable aberration. Thanks for the discussion.

Hardly. Arial is much more limited than Arial Unicode MS.

Ok I should switch to Arial Unicode MS

 

[Animesh K]

Ok here I go.

Fonts which way do you believe it is better to use in html a "%" or a "pt"?

I have has several talks about this with others and without stating my
opinion what would everyone here has to say is better?

Also just out of curiosity what is your favorite fonts to use in a document?

Mine is verdana

Thanks in advance.

Andrew S.

While we are at this ...

Can someone tell which font is closest to the one in "English" in this
PDF document at my website. The pdf was generated using Latex-Dvipdf.
Tahoma seems a bit off.

http://www.stutimandal.com/gif_adi/bhavani_ashtakam.pdf

Thanks in advance,
Animesh

 

[Jukka K. Korpela]

Scripsit Animesh K:

While we are at this ...

I'm afraid this gets far too off-topic; the comp.fonts group would be much
more suitable.

Can someone tell which font is closest to the one in "English" in this
PDF document at my website. The pdf was generated using Latex-Dvipdf.
Tahoma seems a bit off.

http://www.stutimandal.com/gif_adi/bhavani_ashtakam.pdf

I don't quite understand the question. Do you mean the font used in English
texts there? Don't you know what it is? Actually it uses at least _two_
fonts, a serif font for the heading a for the bold texts and a sans-serif
font for the copy text. This violates typographic principles - they tell you
to use just the opposite, serif for copy, sans-serif for headings. Serif
should rarely appear in bold, and inline bold serif inside sans-serif
paragraphs is just... er... very wrong. Some of the characters look awful;
the macron (line above) is barely noticeable in the heading, too bold in the
bold texts and partly misplaced horizontally.

Oh, and an em dash with spaces around is wrong. Use either an em dash with
no spaces (US style) or an en dash with spaces (British style), and make
this match your choice of the version of English.

 

[Athel Cornish-Bowden]

This looks fine, no boxes...

Same for me (also with iCab), but there are some odd browser-related
differences. All three do some font-substitution for the accented
letters in "Totakstakam". In iCab this is done so well that you don't
notice the font is different unless you look carefully, but it's a
lighter font. Second best is Firefox, which also uses a lighter font,
but in this case a much lighter one so it's obvious at first glance.
Safari is the worst, because it uses a serif font in a smaller size
(OK, maybe it's nominally the same size, but it looks smaller to the
eye). In the text underneath the font substitution is almost
undetectable in iCab and Firefox, but very obvious in Safari.

 

[Animesh K]

Scripsit Animesh K:


I'm afraid this gets far too off-topic; the comp.fonts group would be
much more suitable.

I meant the font-family. I found it for myself, it is "Computer Modern
Sans Serif" and I would think it is special to Latex/Tex system.

I didn't follow the english manual of style due to typographic
limitations in the diacritics package and I know why I made certain changes.

 

[Animesh K]

Same for me (also with iCab), but there are some odd browser-related
differences. All three do some font-substitution for the accented
letters in "Totakstakam". In iCab this is done so well that you don't
notice the font is different unless you look carefully, but it's a
lighter font. Second best is Firefox, which also uses a lighter font,
but in this case a much lighter one so it's obvious at first glance.
Safari is the worst, because it uses a serif font in a smaller size (OK,
maybe it's nominally the same size, but it looks smaller to the eye). In
the text underneath the font substitution is almost undetectable in iCab
and Firefox, but very obvious in Safari.

Thanks for the detailed review. I think I will change the font-family to
a diacritics-supporting one in the next round of updates.

Many thanks,
Animesh