Re: integration of general/web in main

[John Mandereau]

Le samedi 15 août 2009 à 09:43 +0200, Jan Nieuwenhuizen a écrit : 
> - And also, please go read the texi2html script and see if you
>    come out alive on the other end.  If I saw someone saying this
>    was great, I might not even want to *see* their printed notation
>    output?

I read it a few months ago, and when I told one of his authors loops
were exceedingly long, which makes the code hard to read, he kind of
laughed at me :-(  I wonder why there are other programming paradigms
like functional or object-oriented besides procedural :-P


> - I would hardly want to advise others to use it.  For me personally
>    it is a maintenance nightmare that I would want to get rid of.
>    25.000 lines of questionable perl.  I cannot imagine that the
>    functionality we need could not be done in 2000 lines of
>    clean, maintainable python

I have been responsible with Karl Berry and Patrice Dumas for deciding
to drop current makeinfo implementation in C about one year ago; we
estimated that writing a Texinfo parser and formatter from scratch would
take years if done by one or a few people as a hobby, so Karl proposed
to adopt texi2html, and Patrice offered to implement Info and XML
backends and improve compatibility with makeinfo.  I don't think we'd
have make another decision today: even if we need a good parser and a
cross-references resolver for docs maintenance, which will have taken a
few dozens of hours to write (it's first time I write a parser, so I
learn what it is as I'm writing and rewriting it), I won't be able to
make a decent formatter for formats other than plain text or Info in
less than a full-time month work, i.e. before December.

Sorry for this kind of off-topic message,
John

signature.asc
Description: Ceci est une partie de message numériquement signée