Re: GNOME font library RFC

[Lauris Kaplinski <lauris ariman ee>]

On Mon, 29 May 2000, Liam Quin wrote:

> Before starting to load fonts across the web, it's important to
> consider copyrights and licences.  Just as the Gnome copyright is
> important, so are font copyrights.
> 
> Most commercial fonts can't be embedded or used on more than five
> CPUs without permission.
> 
> Font vendors *are* interested in ways to make it easier to connect
> font users and their customers, but right now, type designers
> generally sell their fonts for a living.  And they don't get rich :-)

Of course :)

But this applies to sharing all kinds of works over network - you have to
have permission first. But possible copyright infringmenet is not excuse
for not implementing exciting technology, I hope.
On the other hand - should there exist advanced technology for using
high-quality fonts in open-source apps, I bet type designers would be
richer, as need for their work would increase :)

I do not know much about font design business, but as much, as I have
understood, it is the digitizing, which is usually guarded, leaving the
original font designer to its own :(
Or why otherwise published Corel all major computer typography fonts under
different names initially?

> Any scheme to distribute fonts over the net has to address this issue.
> There were attempts to do that with the ISO Font Services mode, but
> I am not aware of any implementations of that, and the standard is
> not freely available.

I strongly argue, this has to be user's responsibility. I do not think,
people would start publishing fonts to www-space more, than they do any
other copyrighted work. And although you CAN get almost every copyrighted
work pirated, most people still really wanting these still buy their
product. At the same time piracy works as massive, low-cost advertising
method for most producers.

Interesting example - the PC/Mac ratio in Estonia is higher than 1000 / 1.
The most important reason for that is, that it is very hard to find
pirated software to Mac here.
Now, if people are slowly getting richer, they are starting to buy more
software - but mainly for PC due to the aforementioned fact.
So - thanks to piracy - M$ has gained a bigger market share here, than it
would gained gained otherwise.

> Having said all that, I am extremely interested in any work to improve
> the dismal state of typography in the X Windows Sysetm and software that
> uses it.  Commercial software like Word Perfect and FrameMaker generally
> has to reimplement client-side font technology from scratch.
> 
> The DSSSL specification has a few notes on internationalisation and
> type that might be of interest, although it's not an easy document to
> read.
> 
> As for fonts being in memory, Most commercial PostScript interpreters
> use a glyph cache -- it's a set of (fontID, pixels-per-em, glyph) triples.
> With a networked font server this could be filled asynchronously too,
> after giving the server a hint about encoding and usage.  With Kanji
> fonts being anywhere from five megabyets to 20 or more, you really
> don't want to have to render the things if yo don't have to.

This was exactly the idea I had, although I did not thought about the huge
sizes of high-quality oriental fonts.

> It's useful to be able to get metrics for glyphs without rendering them,
> too -- a formatter might use the width information for copy-fitting
> without actuallly rendering all of the text.  it's *certainly* useful
> to be able to get metrics for just a few characters.  I wrote a file
> browser once that did this, asking X for a font containing only the
> glyphs it needed, for a huge performance increase.

This is something extremely useful, but I have to think out, how to fit
this into abstract font framework - i.e. how maintain one->many
relationship for X rendering.

Regards,
Lauris