Re: manpage for GNU utilities?
Manoj Srivastava <srivasta@datasync.com> on Sat, Feb 21, 1998 at 10:37:15PM -0600:
> Who needs to pull up emacs to read info pages? there is a
> stand alone info, which is faster than lynx, and has way
> better search capabilities than either lynx or "less".
Info is better suited for certain information storage functions.
Certain programs don't need the nodal, heirarchial nature of info, and
would IMO be better served by a man page; some are on the borderline
(like bash) and some are just obviously better suited to the info format
(like libc). The searching functions in info aren't really
better...it's just that the nature of info pages allows delimiting,
whereas man is a straight line and if you start searching at top, and
don't already know a unique keyword to search for, you'll end up looking
through many bogus hits. Either that, or try to find the appropriate
section yourself, which generally involved skimming until you find it.
Using info, you simply traverse nodes and can arrive at destinations
more quickly.
Info is not without its problems either -- 80-column pages for one; lack
of bold characters; no convenient vi keys, to name a few. There is also
an annoying habit of the top node (dir) to be broken, have duplicate
entries, or obsolete entries.
I think that a man page which says "read the info page instead" is
entirely sufficient; while this may not be the user's preference, the
maintainer nonetheless provides all the necessary documentation; it's
just in a format that some people don't like. But at least the Unix
convention of "man blah" will still hold in that you will be pointed to
a source of more information. And since a Linux/GNU system uses a vast
quantity of GNU stuff -- this is just going to have to be dealt with.
You cannot please everyone.
Alternatively, info could have functionality added, or could front-end
for man pages, or the reverse. Nodal man would be very nice; still
using ?roff macros and such to format output, but using the less pager,
and containing tags and such to traverse nodes. Instead of a central
"top" node (useful for indexing/apropos), there would still be singular
man pages in appropriate man sections. They would contain tags and such
so that the man database updates could handle all the dirty work.
Dunno, I use man-1.5, not man_db :) And info when it's appropriate.
--
Scott
Reply to: