Re: Minutes of the GNOME Board meeting September 10 2002

[<Jim Gettys hp com>]

Roughly, we've been trying to get as much as we can given the financial
constraints (almost no money).

Solving the open source "out of the box" problem is something we must
do to likely get open source desktops to serious volume.  Our presumption is
that people frequently already own a copy of other fonts that are useful to them, and that third party software generally provides good fonts with it.
We feel the most urgent problem to sove is having some presentable fonts that can be truly freely redistributed, so that applications can depend on their
existance, and Gnome (and all other open source software) look good out
of the box.

Once we break into the general market, I expect the financial constraints we face today will become less serious; right now, we have little
credibility with people with deep pockets, something I hope will change
next year as current technologies start serious deployment.  But today,
your first hassle is installation and configuration of *any* fonts at all,
which has left us at a very serious disadvantage relative to Windows and
Mac.  So this is a chicken and egg problem that we must solve to go further.

One of the major font vendors wanted of order $400K US for a really
nice set of fonts: I don't know where we can get that kind of money at the
moment;  there is still too much skepticism in most companies that Linux can
make it on the desktop to venture that kind of money during economic recession
times.  So that option doesn't seem viable right now (it might be once things
start rolling seriously).

We are making serious progress with another vendor, though there have been some rendering issues with the proposed font set (recently resolved), and I18N
coverage issues, for which they will be providing some additional fonts with
wider coverage (this is all presuming the negotiations come to a successful
conclusiong, hopefully quite soon), since Pango/Fontconfig can "do the right
thing" to use multiple fonts at once.  This is not ideal, but would end up with
fonts that are very much better than anything we have available now, with
decent I18N coverage.

Fonts for Chinese, Japanese, etc, are a different kettle of fish and appear
completely out of reach.

Another option, that would cost of order $75K-$100K, would be to get Chuck
Bigelow's Luxi family hinted (it is already freely redistributable
as part of XFree86, but it was developed as Type1 and has never been
hinted well for TrueType, which is why when you use Luxi on your screen
that does not look very nice), but this would also require serious fundraising;
it is probably the option we'll might have to go for if negotiations fall
through.

So, in short, our priority list has been (in this order):
 0) freely redistributable fonts
 1) high quality fonts
 2) I18N coverage
 3) monospace, sans and serif fonts are needed.

Note that I18N is on the list as very important when we've been talking to
vendors.

Nothing else really makes the list particularly (though we do worry about
some nits like distingushing 0O and 1l in the monospace face).

I hope to be able to say more soon.
                                          - Jim

-
Jim Gettys
Cambridge Research Laboratory
HP Labs, Hewlett-Packard Company
Jim Gettys hp com