White space (Was: Graph package, etc.)

[Olaf . Weber]

address@hidden (Jeff Kingston) writes:

[...]

> I see that people are ganging up against me on the matter of white
> space.  Certainly it would not be hard to add a TeX-like space
> munching style to the style information, alongside the font,
> paragraph breaking, and hyphenation style information.  But how
> would you get two spaces between sentences then?  If the answer is
> to define .  to be "." &2s and similarly for ! and ? (quite
> feasible), you get two spaces after Mr. too.  I don't think this is
> a good design at all, I think it's unnecessarily complicated and
> means that the user has to learn things about how to deal with the
> exceptional cases.  Of course, TeX users have already learned these
> things, but the rest of us don't want to.  Prove me wrong about this
> if you can!

> Jeff Kingston

Well, as a TeX user with an interest in Lout, I am of course
prejudiced.  I find TeX's space munching both welcome and overly
complicated.  An advantage of TeX over Lout is that the amount of
work that needs to be done to convert a filled ASCII text is
considerably less, although a fair amount of work is needed in both
cases.  (However, you can always say `\frenchspacing' in TeX, which
sets interword and intersentence spaces to the same size.)

My preference would indeed be an algorithm where `. ', `! ' and `? '
get a larger space than usual, but consistently.  The rule in TeX that
`X. ' would get normal interword spacing is rather awkward.  It is an
exception that in fact, does little good.

As Knuth notes, you do not usually want a linebreak between initials
or `Mr.' and the name that follows.  Hence `Mr.~J.~H.~Kingston' in TeX,
where `~' prevents the break, and also makes sure only an interword
space is used.  This solves both the breaking problem and the spacing
problem for most intersentence periods.

When writing (La)TeX I only rarely need to use `\ ', which in TeX
inserts a single, breakable, interword space.  Its use would be after
abbreviations.

The system that TeX uses allows for spacing rules that are more subtle
(read complicated) than those of Lout.  For instance, a space after a
comma will stretch faster than a normal interword space, but not as
fast as the intersentence space.  And the intersentence space in TeX
is quite diferent from from the double interword space used in Lout.
Only the automation practiced by TeX makes such thing feasible.

Whether non-TeX users are willing to cope with a scheme as sketched
above, or prefer the current rules, I do not know.  It seems that
there is something to be said for both.

-- Olaf Weber