[OS X TeX] Version controls for LaTeX book production

[Mark Eli Kalderon]

On Nov 12 2007, at 08:15, George Gratzer wrote:

> I realize that this is a matter of personal taste. I have been using Retrospect which makes me a few backups a day; anytime I pause, in fact.
>
> My point was simple; this works for me. I have never lost anything and if I needed an older version, it was always there.
>
>I hope it is the same with Time Machine. Nevertheless, I did not stop using Retrospect.
>
> My concern with any tool: efficacy. I am not interested playing with my computer, trying out new things just because they are there. I try anything, however, that promises to make me more efficient.
>
> A Retrospect backup of my current work is one click. Storage capacity: maybe a hundred years. But no project of mine acceded 5 years.

A backup system and a version control system are not the same---I use
both. Subversion, for example gives you an annotated timeline of changes
to a directory with diff, merge, and tagging facilities. Neither
Retrospect nor Time Machine does this.

WRT efficacy, you can manually mimic the functions of a version control
system. (Indeed, the roots of version control lie with engineering,
business, and law where, respectively, bluelining, redlining and
blacklining differences track changes of a document over time---be it a
blueprint, a contract, or some legal document. These practices predate
any software implementation.) So, yes, it is possible to manually track
changes to a document. But, user defined methods are never nearly as
sophisticated as redlining a contract and so are less efficient (in the
sense of providing less information about differences over time). More
importantly, however, *it is always a gain in efficiency to automate
processes done manually*. It's easier, more reliable, and less error
prone.

There is another point about efficacy---it is not just the storing of
revisions, but, as I mentioned above, the diff, merging, and tagging of
differences. Finding differences between revisions manually is utterly
unreliable. So too with merging different versions of a document.

So, without criticizing George for not using version control---he seems
to be both happy and productive with his methods, the suggestion that
people are playing with new fangled software with no gain in efficacy is
simply misleading. (And its not that new fangled---the roots of version
control in software management can be traced back to the 1950s and so
predates TeX and friends.)

All the best, Mark