Re: forwarded message from Jeff Licquia
Walter Landry writes:
> Frank Mittelbach <frank.mittelbach@latex-project.org> wrote:
> > - to get support from the kernel for a new package you have to fork the
> > kernel
> > - when modifying all future names pile up as being unchangeable
> >
> > all of them wrong (and explained over and over again by now)
>
> I must be thick headed. How can you say that the kernel will never
> need to be modified for a new package? I accept that in most cases,
> this is true, but saying that it is always true is absurd.
no its not. perhaps you mistake "kernel" with "virtual machine". LaTeX as
most macro languages is three layered:
1 a virtual machine (TeX program)
2 a kernel (the way LaTeX starts out by default)
3 packages loaded within a document ( document = latex program)
Each package can modify absolutely any aspect of layer 2.
What it can't do is to change 1.
That is: TeX, for example, can't execute OS commands. Then there is no way to
make that possible from (3) or (2).
So to get such an extension you need to modify (1) (which has been done) but
TeX is under a different license.
But comeing back to my statement. The kernel of LaTeX is nothing other than a
huge list of TeX instructions executed and the state of TeX then dumped as an
image for high spead loading at this point.
> Also, why don't future names pile up as unchangeable? The LaTeX
> project release file FOO. Bob modifies and renames it to BAR and puts
> it under the LPPL. Alice modifies that, renames it to BAZ, and puts
> that under the LPPL. Eve modifies it again and doesn't know what to
> name it, since we've run out of silly computer science names.
true, by extending the LaTeX language (through putting BAR under LPPL) Bob has
added to the language and in this way to the pile of names within the
language and so does Alice. This is like extensions in a computer language
from one version to the next might increase the number of keywords in that
language. And yes number of useful keywords is finite.
what i meant, however, (and sorry for not expressing that good enough) is that
LPPL doesn't pile up names by default, ie simply through forking. That is
there is no requirement for Alice to put BAZ under LPPL just because FOO or
BAR was.
frank
--
To UNSUBSCRIBE, email to debian-legal-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: