Re: proposal for inter-word spaces

[Basile STARYNKEVITCH]

Hello Jeff, Hello All,


regarding interword space proposal by Jeff (and related issues), I
have a few comments:

0) I'm not an expert in typography, so take all my comments with
caution! In particular, I'm not sure that "French don't do this
two-space thing". Please check it with reliable sources!

1) I agree with Blake McBride, and disagree with Jeff, in that space
between sentence should not depend on end of lines. Actually, I would
like all space characters (in the C isspace() function sense -ie space
and \t \r \n \f) to be handled as equivalent. The only case where the
end of line character is significant is for @Verbatim stuff - but
since @Verbatim invoke sed with a @Filter it is irrelevant (and by the
way, it should be mentionned that Lout needs sed in the
documentation)..

2) I would like Lout to ignore any extra spaces typed in (except of
course in quoted strings "like this", where they do matter!). I think
that typing one or more blank lines or several spaces should not be
visible on the printed output. I also think that Lout limitation of
input lines up to 256 char is a nuisance (IMHO, the line buffer should
grow as needed), and so is the requirement that a Lout file should end
with a newline.

3) A related issue regarding parsing is that @Filter-ed stuff - or
other portions of text, might contain arbitrary characters up to a
certain magic end string, much like the <<TOKEN in Unix shells, which
could be @<TOKEN. I would like to be able to have in a Lout file a
portion of Html, eg.

@Html @<|--END--|

<H2 CLASS=SC>L'Albatros</H2>
<P CLASS=IL>
Souvent, pour s'amuser, les hommes d'&eacute;quipage<BR>
Prennent des albatros, vastes oiseaux des mers,<BR>
Qui suivent, indolents compagnons de voyage,<BR>
Le navire glissant sur les gouffres amers.
</P>
<H1 CLASS=SC ALIGN=CENTER> ### a line of numbersigns #### </H1>

|--END--|

This would be processed by an Html2Lout @Filter which would read all
of this and output appropriate Lout. The ### and & would be passed to
the filter. I feel that filtered actual arguments lexical rules
(section 2.4 of expert guide) are already too complex! I'm not sure
that filtered stuff should be pre-processed by Lout for @Include-s.


4) perhaps Lout parser should be highly configurable (like TeX),
giving each character a class, but I feel it is too general. On the
other hand, Lout lexical rules (in a parser, computer language, sense)
are already too complex (mostly because of the special handling of the
newline).


5) In my opinion, the only case where newline matters is in poems (ie
versification) but this could be handled much like @Verbatim by a
@Verse stuff invoking a sed filter to add to each non-empty line a
@LineEnd. IF newline had a special meaning, then it should be logical
that empty lines separate paragraphs (like in TeX). Otherwise it
should just be handled as a space character (and also end of
comment). 


6) I don't know how other spacing characters such as tab, verticaltab,
formfeed, backspace etc are handled. Perhaps it should be
documented. (the expert guide mentions tab and formfeed, but not
newline and verticaltab). 

Thanks for reading!

-- 

N.B. Any opinions expressed here are solely mine, and not of my organization.
N.B. Les opinions exprimees ici me sont personnelles et n engagent pas le CEA.

Please cite a **pertinent part** of my mail in all answers
Veuillez citer une **partie pertinente** de mon courrier dans vos reponses



----------------------------------------------------------------------
Basile STARYNKEVITCH   ----  Commissariat a l Energie Atomique (civil)
DRN/DMT/SERMA * C.E. Saclay bat.470 * 91191 GIF/YVETTE CEDEX * France
fax: (33) [1] 69.08.85.68; phone: 69.08.40.66; homephone: 46.65.45.53
email: address@hidden (or else address@hidden);  
I speak french, english, russian. Je parle francais, anglais, russe.
----------------------------------------------------------------------