Re: Social Contract GR's Affect on sarge
Florian Weimer <firstname.lastname@example.org> writes:
> > Once one realizes that, the problem evaporates (and with it, the
> > "current mess").
> Why do you think so? I'd still like to have some hinted non-METAFONT
> DFSG-compliant fonts. And I view the lack thereof an actual problem
> which has bothered me for a long time.
I'm sorry, what's the mess, exactly? The mess so far described on the
mailing list arose from an attempt to apply the GPL's definition to
fonts in Debian.
So I guess really, if there's a mess, can you explain it to me?
The DFSG says "The program must include source code." So for works
which are programs, we know how to apply that.
The new Social Contract makes clear that this is to apply to
non-programs also. But how, since "source code" is a term that
normally refers only to programs? Well, one way would be to take the
GPL's definition of "source code". But we need not do that, and there
are compelling reasons not to.
We should look to the reasons for insisting on source code. One of
those is to preserve the ability of users to modify the program.
Usefully modifying machine binary code for a computer program, except
in the most trivial of cases, is impossible.
So for a program, the requirement to include source code is there to
make it possible to modify the program usefully. Without that, one
would be stuck.
For a font, this is not quite true. Many fonts in Debian are the
output of little languages or the equivalent. So we have no problem
with the METAFONT-generated fonts. IIUC, there is similarly no
problem with Truetype fonts.
There are many fonts which are created as bitmaps. For them, there is
no problem also.
There may be a problem for fonts which were created by METAFONT or the
like, but which are distributed only as bitmaps. Such cases need to
be addressed case by case, and there may be some hard cases where it
is not clear exactly what to do. But we need not use exactly the same
definitions we do for source code, because we should always remember
that digital font designers have historically been quite happy to
modify the output of such programs by tinkering with pixels--that is,
the assumption that one *cannot* modify binaries is true for programs,
but not true for fonts. And, as a result, our judgments about what
constitutes adequate source can vary.
But maybe they need not vary; all I'm saying is that we can look at
the cases and discuss them. The appropriate forum for that is, of