Re: Font license recommendation

[Lars Hellström <Lars.Hellstrom@math.umu.se>]

At 04.42 +0200 2002-07-31, Thomas Bushnell, BSG wrote:
>Lars Hellström  <Lars.Hellstrom@math.umu.se> writes:
>
>> The problem with GPL'ing is that anyone who recieves a PS file using a
>> GPL'ed font could then claim that the PS file in its entirety must be
>> GPL'ed and thus request to get the (.tex or similar) sources for the PS
>> file, since these would be "the preferred form for making modifications".
>
>If the font is really separate: that is, if the encoding is done in
>such a way that it's easily extractable, then it clearly seems like a
>case of a mere aggregation.

It odd to see such a conviction that "this is aggregation, which is
harmless" here on this list, considering that it was recently claimed that
a tarball (!) must be considered to be single work until proof of the
contrary has been obtained, without any objections from the regulars. Can
anyone think of any use other than aggregation for a tarball? But perhaps
there are double standards at work ...

I don't believe that interpretation of the GPL aggregation clause

  In addition, mere aggregation of another work not based on the
  Program with the Program (or with a work based on the Program)
  on a volume of a storage or distribution medium does not bring
  the other work under the scope of this License.

is plausible enough to rely on in my case, but it could be interesting to
examine the matter more closely. Webster's New Illustrated Dictionary
defines "aggregation" as "a collection of particulars" and a "particular"
as "an individual thing or point or quality". Is it then the case that
anything which is a collection of things with individuality becomes an
aggregation? The "it's just an aggregation" argument against the GPL would
then be that as long as you can tell, for each piece of code in a program,
whether it is generated from GPLed source or non-GPLed source, then the
program as a whole is merely an aggregation and the condition in the GPL
that a combined work must be GPLed would not apply.

I doubt this argument could work. However if it did then it certainly would
provide a technical solution to the (obnoxious?) GPL incompatibility
problem: just design the linker so that it pads the executable with markup
saying "beginning/end of material that is part of the work XXX", and then
claim the file is an aggrevation of different works, which just happens to
be interpreted as an executable program by the OS.


Theoretical excursions aside, it seems to me that the Design Science
License (as suggested by Martin Schröder), with an extra clause (as
suggested by Walter Landry) that inclusion in documents is fine, will be
appropriate for what I had in mind. In particular a renaming clause
certainly is a *good* thing when it comes to fonts and the like.

Lars Hellström