[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: man vs info vs texinfo files



On Thu, 2009-03-05 at 15:37 -0800, Per Bothner wrote:
Fedora (and GNU systems in general) should stop shipping info files,
and should instead ship the texinfo source files.  The makeinfo
program will convert texinfo files into html or docbook, as well
as old-fashioned info.  Html or docbook is much nicer - you can
use non-monospace fonts (for the body), and the text can be
re-flowed to fit the window.  Also, indexes and cross-references
can take you to a precise location, rather than just a node.

On Thu, 2009-03-05 at 23:36 -0430, Patrick O'Callaghan wrote:
Interesting idea. Is there any reason why this wouldn't work?

The main problem is someone needs to do the work and drive the
process.

On 03/06/2009 12:26 AM, Ignacio Vazquez-Abrams wrote:
Are we then going to require people to use elinks or w3m to view them,

No.  I guess I wasn't clear.  I was suggesting that one possibility
is to use elinks or w3m *automatically* as backends for the
info program or info mode: Or rather the ideal would be a hybrid
program that can display HTML files, but uses info keybindings and
the info navigation paradigm.

plus expect them to know where the HTML pages are stored in the first
place?

Do we require people to know where info pages or man pages are
located?  Of course not.

The proposal is:
(1) Gnome/KDE help viewers and similar programs should not
display pre-formatted info files, but instead either
(a) format texinfo files on the fly, or
(b) display pre-formatted html or docbook that have
been generated from texinfo.

(2a) (Optionally) To avoid the redundancy of installing both texinfo
and pre-formatted html or docbook one could save disk space
by having info-reading programs cause the makeinfo program be
invoked automatically as needed.

Alternatively:

(2b) Universally replace info format by xhtml, generated from
makeinfo, and modify the info program and the emacs info mode
to be able to read these xhtml files, possibly using elinks
and w3m as a basis.  But this is a larger project - I'd
think focusing on (1) and maybe (2a) would be easier.

No, the far more sane option is to fix info or pinfo so that it catches
SIGWINCH and reflows as necessary.

No, that's not feasible.  The info file format just doesn't contain the
necessary information: It's a more-or-less plain ASCII file, and doesn't
distinguish between hard and soft linefeeds, for example.  Plus you
don't have the information needed to pick the correct fonts.  Cross-
references only identify a named node, rather than a specific paragraph.
All of that information *is* in the texinfo source.
--
	--Per Bothner
per bothner com   http://per.bothner.com/


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]