[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: technical question

> Would it be possible, using the existant macro in plain to produce an
> estensible integral sign ?
> 
> on a font and charlist point of view, there is no problem !
> But i'm wondering if the \left\int bit would work.
> how should \int be defined ?

Is it possible, that in a paper where \left\int is used, it will be
desired to have \int always act as a delimiter, and never in the plain
TeX way? If so, changing the definition and syntax of \int might
be the best approach, so that the \left is built-in. For example:

  \def\int#1#2{\left\intdelim #1\right.#2}

with usage:

  \int{f(x)}{dx}

(I don't recall from previous mail, is the differential placed after
the \right. or before?)

Of course there are complications with subscripts and superscripts
that would have to be dealt with.

As a syntax of this sort could also handle non-delimiter integrals,
the obvious next thought is a question: whether the old integral
syntax should be retired in favor of a new syntax. The old
backward-compatibility viper rears its ugly head. Using a new name
e.g. \integral would help, but at the cost of taking longer to type.

As for backward compatibility in general: Of course it is very
important, but what about the practical consequences? Ask Joe User if
he wants 100% backward compatibility for his existing ten or 100
documents and he is sure to say yes; but next ask if he is willing to
wait two extra years for the new math font encoding because it will
take that long to iron out the problems of achieving 100% backward
compatibility. Or, in other words, the benefit to the end-user of
striving for perfect compatibility is not so clear-cut if you factor
in the disservice to the end-user of delays for new software releases.

Michael Downes                              mjd@math.ams.org (Internet)