[tex-live] Re: [tex-fonts] error in ec encoding

Lars Hellström lars.hellstrom at residenset.net
Thu Aug 18 18:40:59 CEST 2005


At 17.52 +0200 2005-08-18, Nelson H. F. Beebe wrote:
>It is now more than 13 years since the platform-independent checksum
>utility and standard file-header comment support for GNU emacs were
>developed after a late-night brainstorming session by a large group of
>us in Heidelberg, and then later implemented by Robert Solovay
>(Berkeley) and me (Utah):
>
>	ftp://ftp.math.utah.edu/pub/emacs
>	ftp://ftp.math.utah.edu/pub/checksum
>
>	http://www.math.utah.edu/pub/emacs
>	http://www.math.utah.edu/pub/checksum
>
>They should be used for ALL of the files in the TeXlive and CTAN
>archives, so that such files carry author information, a version
>number, a timestamp, a checksum, and other documentation in a standard
>form, making it easy to figure out which version is current, who is
>responsible for it, and for end users to verify whether the file has
>been modified without a header update since it was released.

WOW! Documentation!!

For years now I have been mystified by these pseudo-BibTeX (do I detect
also a hint of MFT in the choice of %%% as line prefix?) headers some TeX
files carry. (At one point I even hacked together a script of my own to
update the timestamps in the fontinst files which carry these headers,
since an incorrect such header is far worse than no header at all.) This is
however the first mention I have seen _anywhere_ of what created these
headers in the first place.

IMNSHO this indicates that the "standard" hasn't been very widely
publicised, and thus is perhaps isn't that strange that is isn't being
followed.

My initial enthusiasm is also somewhat quenched when I look at the things
these URLs point to. Rather than being a documentation/specification of the
format, the material appears to be software. This is perfectly fine if you
can immediately install and run that software, but from a superficial look
at it this seems far from certain. Maybe there is a specification somewhere
in one of the tarballs, but right now I don't have the time to explore that
further.

An Emacs solution is not a universal solution.

>Once the EC.enc mess is sorted out and the definitive version
>identified, let's add a standard header to it, put it in a version
>control system, and track the changes to it that way.
>
>A 500GB SATA disk
>
>http://www.hitachigst.com/portal/site/en/menuitem.8f07a3c3d3a7a12d92b86b31bac4f
>0a0/
>
>now lists for US$437; we can easily have the resources to store
>complete revision histories of every file in the TeXlive and CTAN
>archives.

Is this a suggestion to extend CTAN with CVS capabilities? Good idea!


>Please remember that one of the great strengths of TeX is its
>stability over decades.

Hear, hear!

>It is important at many sites to be able to
>retypeset documents that were written years ago, and to do that, all
>of the style files used by the documents must be available.  That is
>why the revision control system either needs to be publicly
>accessible, or else version-numbered packages of every collection of
>files need to be available back to the day they were first released.
>
>While single users with personally-owned machines may enjoy always
>having the latest-and-greatest versions of every piece of software,
>corporate and institutional realities require much more stability and
>reliability.

Amen.

Lars Hellström




More information about the tex-live mailing list