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

Re: Encoding the name in the file contents

Lars Hellström  <Lars.Hellstrom@math.umu.se> wrote:
> At Thu, 25 Jul 2002 12:50:49 -0700 (PDT), Walter Landry <wlandry@ucsd.edu>
> wrote:
> >Boris Veytsman <borisv@lk.net> wrote:
> >> Let me tell you how the things are organized in the TeX world. There
> >> are dozens of TeX implementations. Some are free, some are commercial,
> >> some are open, some are closed. I would not be surprised if some of
> >> these are not written in C and do not use the standard web2c. The
> >> authors of these implementations package LaTeX with their
> >> systems.
> >
> >I see now.  I was concerning myself with free software, whereas you
> >have a desire to interoperate with proprietary software.  Fair enough.
> >What if this md5sum were computed using TeX?  Assuming reasonable
> >performance, would that be a solution?
> It would still be pointless, since there would be nothing to compare the
> computed checksum against, unless the file itself says what checksum it is
> supposed to have, and in that case we're just doing a more complicated
> variant of the \NeedsTeXFormat.

Um, no.  In the case where package FOO needs package BAR,
\NeedsTeXFormat has BAR tell FOO that BAR is a good version.  Using
checksums, FOO checks for itself whether BAR is a good version.  That
way, if anything changes in the chain of packages, it propagates up to
the LaTeX kernel because of mathematics, not laws.

> Furthermore you most likely couldn't get anything near a reasonable
> performance for an md5sum implemented using TeX macros. There are, for
> example, no bit operations available in the language. If you think you
> would need any, then you would have to tabulate them, and that gets
> unreasonably expensive (about 20 times the "cost" of the LaTeX kernel and a
> couple of packages) already for byte operations.

It wouldn't have to be MD5 per se.  It could be something more easily
computed that is good enough.  This is only to prevent casual
fiddling.  A determined person would find it much easier to do some
macro wizardy to fool the system.

> >Boris Veytsman <borisv@lk.net> wrote:
> >> Even if this were possible, did you consider the logistics nightmare of
> >> conversion of megabytes of LaTeX code written by hunderds of authors
> >> to the new authentificaiton scheme?
> >
> >Wouldn't this logistics nightmare happen with a license change as
> >well?
> Why should there be any need to change anything but the text of the
> license? That's only one file.

The hard part is getting the person who owns the copyright to agree to
the change.  Actually changing \usepackage to \use_and_verify_package
is not.  Remember, it doesn't have to be done if the copyright holder
doesn't care.  And most people don't care.

Walter Landry

To UNSUBSCRIBE, email to debian-legal-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Reply to: