[texhax] Today's binaries on old texmf trees

William F Hammond hmwlfsr at yahoo.com
Tue Nov 5 02:33:24 CET 2013

Reinhard Kotucha <reinhard.kotucha at web.de> writes:

> On 2013-11-03 at 13:47:56 -0800, Alex Scorpan wrote:
>  > In an attempt at developing a long-term preservation strategy for
>  > old source,
. . .
> Alex, I strongly recommend users to avoid such setups.
. . .
> Nevertheless, if you are interested in long-term
> preservation, the better strategy is to test your sources
> on recent systems and inform package authors if something
> doesn't work as before.  I'm convinced that in the long
> term this approach is more reliable.

I take this to be in the context of mathematical journal
articles.  Most authors, who generate the sources, are not
interested in the finer details of typesetting, and many
will be irritated to receive such a report from a publisher
five years after publication.

It won't be a retrospective solution for preservation, but
for the future I believe a better way will be to seek
formally profiled LaTeX source from authors.  Such source
will not only admit automatic translation to online formats
but also translation, independent of any particular TeX
engine, to the LaTeX flavor of the day.  Of course, as time
goes by, not only will LaTeX evolve, but also the LaTeX
profiles will evolve, so there will be a need to maintain
libraries of translation software keyed to both source and
target versions (which could -- really _should_ -- be a
community project).  (If profiles evolve so that old
profiles can be translated to new, that would save some

                              -- Bill

More information about the texhax mailing list