[texhax] Obsolete \centerline command used in amsbook class (Uwe L?ck)

[Barbara Beeton]

On Sun, 23 Sep 2012, Michael Barr wrote:

    I would like to comment on Uwe Lueck's comments on the use of plain and
    primitive tex commands in a latex document.  [...]
    he felt about that (this was at least 20 years ago) and his reply was, in
    effect, why not.

    [...]  One thing that really irritates
    me about amstex is that they have undefined some tex primitives.  One is
    \over (and its relatives).  [...]

please, michael, amstex and ams-latex are *not*
the same thing.  too many people think they
are, and it causes no end of problems in ams
production -- which makes producing journals
more expensive.

    [...]

    And don't get me started on how much more functional \def is than
    \newcommand.  I authored some diagram-drawing macros (a front end to xy-pic,
    actually) some years ago and they would have been impossible to do using
    \newcommand.  But when I use \def, someone is sure to comment that that is
    not latex.  No, it is not, but when did latex turn into a religion?

when authors know what they're doing, and
use \def only for command names that aren't
already defined in the enviroment they're
using, it's okay.  but redefining primitives
is fatal, and last week we received a paper
that redefined \newcommand!  this wasn't
caught, and happened to work until the last
stage of production, in which dual versions
are generated, one for paper without links,
and one for on-line, with live links, along
with an (automatically derived) html version
of the top matter and references for direct
browser viewing.  then it crashed, sending
the file back to square one, and delaying
delivery of the journal issue to the printer.

    Which is not to say that latex doesn't provide lots of functionality that
    plain doesn't.  But it remains, like the underlying tex engine, just a tool.

there are good reasons to request that authors
avoid the use of \def.  contemporary production
and delivery methods require uniformity, and the
least expensive way to accomplish that is for
the authors to apply the tools in the cleanest
way that gives the most reliable results.
						-- bb