[texhax] Latex: dumbing down? (fwd)

Barbara Beeton bnb at ams.org
Fri Aug 25 18:02:02 CEST 2006


    I know this post will start a flame war, but I have no intention of 
    getting involved in that or answering it.

i have no intention of starting a flame war,
so we're even.

    I recently got a paper to copy edit that contained the line
    \[ (*)\ \ \ \ \ \ \ \ \ some formula \]
    with a later reference to formula $(*)$.  Naturally, I changed that to
    $$ some formula\eqno(*)$$
    (Actually, first I tried \[...\eqno(*)\] but that gave an error message.)

    I began to muse about this.  I have been lectured by latex purists about 
    how dangerous it can be to mix plain and latex (why then is nearly the 
    entire body of plain incorporated into latex?), but I have never paid the 
    least attention to that.  Certainly Leslie Lamport was not rigid on that 
    question.  Two pages of his "Latex" was devoted to that question.  As far 
    as I can tell from that book as well as the Companion, there is no way, in 
    pure latex, to put in (*) as an equation "number".  But this is a 
    perfectly reasonable thing to do.

if you look at the amsmath user's guide, page 3,
you will see that it's definitely possible to
have (*) as an equation number.  you just have
to use the appropriate package.

    More generally, restricting oneself to "pure" latex, while a good idea for 
    a beginner, is too restrictive for someone who wants to get the most out 
    of tex.  I am beginning to understand the mentality of a colleague of mine 
    who told me that he would never use Latex because Latex does not allow you 
    to...(whatever, as a matter of fact, I think it was a matter of using 
    \eqno).  I told him that latex incorporates nearly all of plain tex, but I 
    could never convince that the things that latex did well (sectioning, 
    lists, etc) were real time savers.  

    Occasionally (although not for several years) the journal will get a paper 
    in plain.  My first reaction is to add our document class and begin and 
    end document and see if it runs.  It usually does.  Of course then I have 
    to go through the paper to replace every theorem, proposition, lemma, 
    etc., by the appropriate latex macro.  Once upon a time, the journal tried 
    to maintain a plain style file, but it was too hard since every user of 
    plain was accustomed to rolling his own.  Now if it comes in in Latex 
    (even if it is just latex 2), I can just make a few changes and it works.

and here you see the main benefit of using latex
with an appropriate package.  a submission for
publication in a journal, if prepared according
to the publisher's guidelines, will zip right
through with no problems -- a tremendous saving
of effort and cost.

    Incidentally, one of the main features of tex (not even plain) that I use 
    heavily for my own work is the ability to use paramater contexts.  As an 
    example,
    \def\comb(#1 #2){\pmatrix{#1\cr#2}}
    allows me to type the very natural \comb(3 2) get what I want.  You cannot 
    do that in pure latex, although it works perfectly in latex anyway.

it may work perfectly for you, but in our journal
production, it fouls things up magnificently, and
someone has to go in and make possibly extensive
changes -- leading to the possible introduction of
errors.

the reason for this is that our journal files are
processed both for print and for on-line posting,
and the problems come with the use of latex2html
to re-purpose the file (only the top matter and
bibliography; the math content is just too horrid
to contemplate), and \def alone (even without the
context marking) causes errors in the output.
i have learned that latex2html has a way around this,
but our procedures were created long ago, possibly
before that mechanism was available, and changing
these procedures to allow the use of that mechanism
would be a very major task.

but we have also discovered a nasty hack that will
allow such coding to remain:
in the preamble:
  \newcommand{\myownmacros}{%
    \def\comb(##1 ##2){\pmatrix{##1\cr##2}}
and invoke \myownmacros somewhere after
\begin{document}, or, even better, \maketitle.

i'm still a plain tex user by preference (there
are some things for which the latex overhead is
just too great, and for which too much has to be
undone), but for the production of journals and
other "mass" processing, latex, properly used,
can cut down significantly on time spent, problems
encountered, and cost.  for that, we put our effort
in up front in developing robust, flexible author
packages, and just hope that authors will use them
as directed.
							-- bb


More information about the texhax mailing list