[texhax] About MathTex

[bnb at ams.org]

hi, bill,

    You write in part:

    > however, the "style" of latex coding is not
    > the best.  unfortunately, many of the examples
    > shown on sites with such material neither
    > exhibit "best practices" nor have undergone
    > validation or editing.

    I don't understand what you mean by 'validation'
    in this context.  Is there something more than running
    it through, say, pdflatex, checking the appearance of
    the output, and checking for noise in the log?

validation:
 - does it run without errors?
 - does it produce the intended
   output?
 - does it avoid redefining
   primitives or commands that
   are defined in basic latex and
   ubiquitous packages (such as
   hyperref)?
 - is the log free of warnings
   about things that aren't
   obvious in the output?
(of course, things like overfull
boxes will depend on the document
class ultimately used; two-column
pages have less leeway than do
one-column pages.  a reasonable
test would *not* assume that
the full width of a4 paper is
available, but would use the
dimension of a typical journal.)

editing:
 - does it show use of "best
   practices"
   - using \mathbf instead of \bf
   - putting braces not just where
     absolutely needed, but also
     where they will make it easier
     for "derivative" processing
     (such as conversion to mathml)
     to be bulletproof;
   - avoiding eqnarray;
   - defining (with \newcommand,
     not \def) commands for common
     notation that may be changed
     during the course of writing?
 - is it laid out cleanly so that
   someone reading the input code
   can understand it readily?

    (Of course, you probably know that I think there should
    be something more ...)

i surely may have missed something
that you would find necessary, or
at least desirable.  please feel
free to add to this list!  it's
good to get this information out.
					-- bb