Fedora 7 CD Labels & Covers

Nicolas Mailhot nicolas.mailhot at laposte.net
Sat May 26 10:19:08 UTC 2007

Le vendredi 25 mai 2007 à 19:08 -0400, Máirí­n Duffy a écrit :
> Hi Nicolas,
> Nicolas Mailhot wrote:
> > Le vendredi 25 mai 2007 à 13:38 -0400, Matthias Clasen a écrit :
> >> On Fri, 2007-05-25 at 17:09 +0200, Nicolas Mailhot wrote:
> >>> FYI the liberation fonts are nice and it's certainly worth showcasing
> >>> them but they're not Fedora 7 default fonts, since they lack extended
> >>> glyph coverage and lack hinting.
> >> Yes, so what ?
> > 
> > So using them systematically is not reflecting the project state
> > accurately, and it will hurt the local Fedora groups when they try to
> > localise Fedora marketing material and discover their language is just
> > not supported.
> I hope it will help to give the reasons why I chose it. I will most 
> certainly switch it to DejaVu Sans, which will have better i18n support, 
> right? I honestly didn't realize Liberation Sans had a coverage problem; 
> I didn't try any non-basic English glyphs with it so that's my bad.

Liberation Sans has not a coverage "problem", it already includes greek
and cyrillic glyphs (which is more than Bitstream Vera when it was
released, or most latin-only fonts you can buy). It's just that it does
not really compare to some of our other fonts, and even its coverage of
latin, greek ans cyrillic is not complete (so big languages will work,
minority/local languages won't).

A very crude way to assess a font is to upload it on
http://www.fileformat.info/info/unicode/font/custom.htm read the glyph
count and look at the glyph matrix. A much better indicator would be the
language support coverage matrix DejaVu publishes with every release but
its generator needs fontconfig sources and fonts in sfd format (IIRC) so
it's not useful for the average artists. We really need a tester tool
that would do the same on any on-disk ttf/otf font.

Anyway, some result for F7 common FLOSS fonts:
- DejaVu Sans/Serif: 4538/1939
- DejaVU LGC Sans/Serif: 3532/1881
- Linux Libertine: 2274
- Gentium: 1699
- Liberation Sans/Serif: 668/661
- Vera Sans/Serif: 268

And another we should really get into Fedora, but has a build system
from hell (and what's the point of FLOSS if you can't work from sources)
and is in opentype OTF format all out tools can not handle yet: Computer
Modern Unicode
( http://canopus.iacp.dvo.ru/%7Epanov/cm-unicode/ )

> (As an aside, how is URW U001 [1] in terms of glyph coverage? Is that 
> another option?)

IIRC URW U001 is free-to-use but not especially FLOSS, limited to latin
(and even basic latin), and in legacy type one format people are moving
away from.


> I could have used DejaVu Sans, but to be honest I'm really sick of Deja 
> Vu's look since it's one of the only good free & open fonts out there 
> and I've seriously used it to death! :)

You have to compromise.
Entities with an international reach (like Fedora) almost never use the
very distinctive fonts you way be used to as a designer, for the
following reasons:
- it's very hard and costly to create international fonts
- it's even harder to decline an original font
- any substitution for coverage reasons is going to stand out. And
projecting a common worldwide style is more valuable than being cool in
a few countries
- very distinctive fonts are distinctive because they're different from
the ones people are used to read, ie harder to read.
- Title fonts in particular are useless for plain text, which is another
reason people don't bother extending their coverage
- very distinctive fonts often do not age gracefully (trends change, a
style may be great one year and laughable 5 years later)
- periodic re-styling is no option, it's overly costly and hurts the
common unified persona you want to project

It's probably OK to use Liberation for titles. The benefits of
showcasing it may outweight the fact it won't be a choice for many
languages. Using it for text where the glyphs are too small for most
people to see the difference OTOH is stupid IMHO.

Also only using Liberation fonts (as I've seen proposed, including in
the message I reacted to) is sending a message I don't agree with (as
already explained)

> How are non-technical Fedora groups making Liberation a Fedora emblem? 

The original RH PR release wrote about FLOSS fonts intended to replace
[a lot of things] (when it will be finished), and many people have taken
it as "Fedora intent is to use Liberation in all its documents *now*
because it's the only realistic FLOSS font"

> But Bitstream is a professional foundry that has closed fonts as well, 
> no?

Current default is DejaVu LGC which is not produced by Bitstream but by
a FLOSS project. Also there's no possible comparison between the
handling by GNOME of the Bitstream fonts and the handling by Red Hat of

> Or did Bitstream do the work and then license it openly? 
> (It's a real question, I actually honestly don't know; I was pretty
> sure the latter was the case though.)

Bitstream did most of the work but the final stage saw a dialog between
the designer and the FLOSS community (feedback, change demands, etc.).
Vera was not just a PR release with some files to download.

http://www.gnome.org/fonts/ has archived all of it.

Nicolas Mailhot
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: Ceci est une partie de message num?riquement sign?e
URL: <http://listman.redhat.com/archives/fedora-art-list/attachments/20070526/8f8b7311/attachment.sig>

More information about the Fedora-art-list mailing list