[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: TeX Licenses & teTeX

>>>>>>> On Tue, 6 Aug 2002 08:08:52 -0600 (MDT), "Joe Moore"
>>>>>>> <joemoore@iegrec.org> said:
>>> Thomas Bushnell, BSG wrote:
>>>> But can I modify the behavior of any part of LaTeX, including what
>>>> happens when I load article.sty?
>>> Yes.  But in order to do so, you either have to:
>>> 1) request such a change in your document (\documentclass{myarticle},
>>> or \renewcommand\documentclass or something like that)
>> That doesn't meet the stated goal, of silently changing what happens
>> when \documentclass{article} is used to load a file called article.sty.

> Yes, it does (literally)  Now that I think about it, it meets one of the
> LaTeX project's goals (use of non-standard LaTeX requires a conscious
> decision), but not Thomas's.

I don't agree with you here, but I'm not entirely certain in my
opinion.  It appears that this would load a file myarticle.sty in the
first case (\documentclass{myarticle}) and not achieve the goal of
changing what happens whenever article.sty is loaded in either case.
Am I misunderstanding the method suggested?

(And isn't it article.cls, anyway?)

> It does, though, since LaTeX is interpreted, and can override any aspect of
> article.sty at any time.

> The first parenthetical option (change the name to myarticle) I would
> classify as an API change, which would be a non-free _requirement_.  The
> second I would classify as an "excessive burden", which would be a non-free
> _requirement_.

>>> or
>>> 2) Modify latex.ltx so that it looks for myarticle.sty when a document
>>> asks for article.sty (but since latex.ltx is covered by the LPPL, you
>>> would have to change the name of it, invoking it by a different name
>>> such as notLaTeX)
>> Again, that doesn't meet the stated goal, since it requires a file
>> myarticle.sty and doesn't actually load article.sty.

> Right.  But it's only one of the options.  By itself, this would probably be
> a non-free _requirement_.

Yes, but it's not even one of the options available to achieve the
goal of "change what happens when the file on disk called article.sty
is loaded."  There are plenty of ways of achieving the goal "typeset
this document differently" or "provide a standard way of typesetting

>>> or
>>> 3) Place your modified article.sty in a different location, where
>>> "pristine" LaTeX will not pick it up and behave silently different.
>> That's a requirement which restricts the functional modifications
>> which may be made, and additionally does not appear in any LPPL draft
>> I've seen.  I'd *love* to get a look at the latest draft, since I
>> think it might clear up some of the confusion on both sides.

> I don't think it makes any functional restrictions.  Only that it has to be
> located outside of the default search path.

Since I am restricted from then copying the pristene LaTeX into
someplace where it will find my modified article.sty, and I am also
restricted from changing the search path of my copy of pristine LaTeX,
it is a functional restriction in how I configure and use my copy of
the software.

> Or combine 2) and 3) and make the default search path of notLaTeX include
> your modifications.

>> Oh, and you forgot
>> 4) Create a new article.sty and drop it into place.  If what's been
>> said about each representation in LaTeX code having a wholly
>> different meaning, creative expression in the language is
>> impossible, and you could crib a great deal from article.sty (which
>> Frank's asserted is purely functional, and not a creative work)
>> without violating any copyrights.

> I'm not sure I understand your representation of Frank's assertion there, or
> how it impacts copyrights.  Is there some part of copyright law that
> restricts applicability for functional works?

Yes.  You cannot copyright an idea, only the expression of an idea.
So Knuth could have copyrighted tex.web, but not the idea of any
system which performed that transformation.  GCC is copyrighted, but
that copyright doesn't cover 100% compatible C compilers.

For this reason (in combination with others), a bitmapped font is not
copyrightable: that representation is the *only* representation of
that idea, and copyright can not be construed to grant protection over
an idea (that's the domain of patents).

Frank M. asserted that any change to the code of the LaTeX system
would change the output.  If that assertion is correct (I don't
think it is, though enough LaTeX developers claiming it would convince
me), then LaTeX is the only conceivable expression of that idea.
Therefore it is purely functional, and uncopyrightable.

As I said, I don't actually believe LaTeX to be uncopyrightable, and I
don't believe Frank thinks it is either -- but that's because I don't
believe (and don't think he believes) that *any* change to code in the
LaTeX source tree *necessarily* changes output.  Consider the addition
of comments, or change in variable names, or the choice of a
particular sorting algorithm over a slower (but equally correct)
alternative, for example.

I think that a great deal of this would be cleared up by the
presentation of a new LPPL draft.


Attachment: pgp22tsUT8CyX.pgp
Description: PGP signature

Reply to: