Re: Lout parsing

[Ian Jackson]

Peter da Silva writes ("Re: Lout parsing"):
> What's wrong with using different symbols for the two different kinds of
> period? It's not so often that you use a dot other than at the end of a
> sentence, why not \. it?

The whole problem with TeX's way of doing things was that people don't
like having to type `weird' things for perfectly normal situations,
like abbreviations.

IMO there should be a sensible default behaviour, along the lines of
checking whether the spaces are doubled or the punctuation appears at
the end of the line.  There has to be a way to override this default
so that certain problems can be resolved.  (Whole sentences that are
parenthesised like this one come to mind as a possible example.)  This
overriding syntax can have @'s (and \'s, if we must) in it.

The default behaviour should be:
  Sentence-ending punctuation at end of line, or doubled spaces in the
   middle of a line turns the whitespace there into a sentence-ending
   space.
  All whitespace is compressed to one space, non-sentence-ending
   unless it is sentence-ending according to the rule above.
  There are two special symbols which force the space surrounding them
   to be regarded as sentence-ending or non-sentence-ending.

You additionally need the following global configuration options:
  Set the size of the sentence-ending and non-sentence-ending spaces.
  Turn off the compression of whitespace.
  Force all spaces to be sentence-ending or non-sentence-ending unless
   explicitly forced using the special markup.

This functionality can be made to do what everybody wants, I think.

Ian.