Re: User's thoughts about LPPL
Nick Phillips writes:
> On Fri, Jul 19, 2002 at 04:04:33PM +0100, David Carlisle wrote:
> > It is not reasonable that the author of a package such as
> > "indentfirst.sty"
> > for example (which consists of exactly 4 TeX tokens) should be expected
> > to go to the trouble of trying to legally register that name.
>
> No, agreed, but...
I think what needs to be added here is that indentfirst.sty is one of the
"extreme" cases. At the other end of the spectrum (also extreme) are very huge
packages. But both extremes are extremes. the majority of packages are in the
middle and it seems hopeless to think that for the
frank@istrati:~ > locate /usr/TeX/texmf/tex/*.sty | wc -l
1375
it seems hopeless to think that registered names would be possible. (.sty =
latex packages to be loaded by \usepackage)
> > the LPPL as it is currently drafted gives other users some assurance
> > that if their documents have
> > \usepackage{indentfirst}
> > then their documents will behave as they expect (indenting the first
> > paragraph of sections) and not invoke some other completely different
> > code that someone thought was an "improvement".
>
> This falls down because the "indentfirst" package is likely to be a relatively
> small work, and until and unless it is incorporated into the LaTeX
> distribution, there is absolutely no assurance that someone else somewhere
> else won't come up with something with the same name, similar intent, but
> critically-different results, without having heard of the first package.
actually it is part of the tools-distribution of LaTeX but i don't see that
this is any guarantee either
> In short, the protection of which you speak is only really present when the
> package is included in a distribution large and widespread enough to justify
> a trademark in any case.
>
> Presumably CTAN will only provide one package with any given name, and this
> fact alone is enough to provide the degree of protection required for such
> a package, is it not?
It is not. The combination of three things make it work
a) LPPL
b) literate programming conventions within the LaTeX community
c) CTAN
why is that so?
1) if the person knows about indentfirst he kows that he is not allowed to
modify it without giving it a new name. he is of course allowed to have
some completely other work given that name, but he will know that CTAN is
going to suggest to give it a different name (so that latex user can use
both packages) --- he is of course free to distribute it from some other
archive.
2) if he doesn't know about it, he might start in good faith with his
"indentfirst". if he uses dtx as documentation system (as most people do)
then when turning .dtx into .sty latex will warn him that it already knows
some indentfirst.sty. This will bring us back to 1) early in the game. of
course there is no guarantee for this whatsoever so he might find out only
later when submitting to CTAN.
The practice has shown that this combination works.
CTAN conventions alone are not enough. They would keep CTAN sane but they
would not ensure that major or smaller distributions of LaTeX start to add
their own changes without renaming the involved files. They are entitled to
make such changes and in fact encouraged to do so, but not in a
way that the changed packages appear from within a LaTeX system ie when
running a document requesting the original package as being that original.
The problem is that in 99% of the cases the wish to make a modification is not
due to a clearcut bug fix but involves an aesthetic component. Note that we do
not argue the validity of a wanted change. we argue that it is best if that
change is offered identically to everybody and by requiring a name change this
is automatically happening as then if that changed package is used anywhere in
a document by somebody the situation is either
- it compiles at a certain site because the changed package is installed
there, or
- one gets the information that package X doesn't seem to be installed
first is perfect, second is good and at least much better than haveing the
document behave differently in subtle or less subtle ways.
Note that there is no question of quality involved. The individual quality of
varioref versus varioref2 might be better/worse/uncomparable. The importance
is that it is either identically available at all sites when running LaTeX or
absent, but not slightly different.
> If there are a large number of third parties using the LPPL, all the more
of the 1375 packages used as an example above 416 directly refer to LPPL and
at at least 400 hundred more are style files belonging to bigger packages
where the whole package is put under LPPL. Then there are a number of styles
which have (if you like a simplified LPPL) lines like
%%% copyright = "Copyright 1995, 2000 American Mathematical Society,
%%% all rights reserved. Copying of this file is
%%% authorized only if either:
%%% (1) you make absolutely no changes to your copy,
%%% including name; OR
%%% (2) if you do make changes, you first rename it
%%% to some other name.",
in them.
so all in all definitely more than 800 .sty files are under LPPL of which less
than 10% are written or in any way influenced by the LaTeX Project Team.
> reason to keep it simple, IMHO. I would at least move the sections regarding
> maintainership into a separate document, and vigourously attack the rest
> of the draft to remove unnecessary complication.
well, the suggestions/request for adding a maintainership part came out of
that 90% of people using LPPL.
> While I'm at it, thanks for spending the time both on LaTeX, and such
> potentially head-banging-against-brick-wall discussions as this.
>
> We appreciate it.
thanks for saying
cheers
frank
--
To UNSUBSCRIBE, email to debian-legal-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: