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

Re: preliminary EuroTeX paper

Matthias Clasen <clasen@pong.mathematik.uni-freiburg.de> writes:

> Another question for discussion: Is it really a good idea to have the 
> double accents in MSP ? They will be missing in most implementations.
> ... Even if we decide keep the double accents somewhere 
> (eg in MS1)

That sounds like a good idea to me.

> ... we might rethink the choice of combinations: I would expect
> the double hat to be the most frequently used double accent, and it is
> not currently among the combinations in MSP. Does the AMS have statistics 
> about the relative importance of double accents ?

I looked up some statistics that I compiled a few years ago when the
predominant macro package for our documents was AMS-TeX (more than 95%).
The sample size (328 documents) was I think too small for the results to
be as useful as one might hope---for example, \Hat was indeed the most
commonly used double-accent macro but it only occurred in one of the
documents (30 instances in that document). By comparison \widehat and
\hat occurred in 71 and 63 documents, with similar stats for \widetilde
and \tilde. In compiling the stats, however, deprecated usage like
\hat{\hat{...}} would not have been caught, since the filter only
counted control sequences, not combinations.

I could mail you the stats file separately, if you want to see it anyway
in spite of its old age, small sample size, etc. It's about 600 lines so
I hesitate to post it to the whole list.

I have an unfinished `symfreq' package that counts symbol usage and
writes symbol stats to a file when you run LaTeX on a document; it is
layered on top of the flexisym package that comes with breqn, since that
provides the necessary hooks to attach arbitrary operations to math
symbols. Let me see if I can find some time in the next few weeks to get
that working better and make it available.

Michael Downes