For what it's worth, my own conclusions about an XPL for having
two spaces after a sentence follow. And one could say it's in
two parts.
First, why the program? Well, it was interesting to say the
least to experiment in different aspects of Xy 3+ and 4. Also,
in terms of traditional publishing, text spacing is either
regular or french, the latter employing the standard word
space. DTP, as we know it, generally only incorporates french
spacing; I can't recall seeing regular spacing specified via
DTP any too recently, but although the XPL incorporates a
double-space, it could be changed to a space plus en space
easily enough (for regular). So oddly enough, it does have a
purpose.
Second, Xy 3+ is still plenty powerful. The second versions of
the two-space program were fairly similar, because the way
defines and mistakes are recognized seem to be a tad different.
The third version (for 4) is the fastest, but Xy 3+ isn't far
behind. The test file was about 64K and there were 1083 changes.
---------------------------------------
4 timings 3+ timings
---------------------------------------
166 sec (with BCs) 35 sec
40 sec (with BXs)
26 sec (with BXs, shortened
version 3)
---------------------------------------
So while Xy 4 is the overall winner, it is also a lot easier to
construct a slower program (by over a factor of 4). I suspect
the XS to extract strings slows the program, but that's only a
guess.
In any case, 3+ still remains a remarkable program.
--Chet
--- þ SLMR 2.1a þ Art + write + dtp = chet.gottfried@asb.com