Re: Custom undocumented(7)s are just as bad.
On Sun, Jan 30, 2000 at 02:18:56PM -0600, Steve Greenland wrote:
> On 30-Jan-00, 08:53 (CST), Michael Stone <mstone@itri.loyola.edu> wrote:
> > On Sat, Jan 29, 2000 at 10:18:18PM -0600, Steve Greenland wrote:
> > > I'd much rather have useful info in README.Debian: this is what you need
> > > to do to finish configuring (if necessary), here's a one-liner for each
> > > major binary of the package, here's what to read to find out more (info
> > > pages, man pages, web site, whatever).
> >
> > As opposed to a man page for the binary that could be contributed back
> > upstream to be useful on systems besides debian?
>
> If I'm only going to get one or the other, yes. It was based on the "we
> need man pages for each and every binary in the package so that people
> can figure out what the package does" reasoning -- for that purpose, I
> think a single file with brief description would be more useful than 20
> man pages.
It's not so you can figure out what the package does, it's so you can
figure out what each binary in the package does. A brief description
doesn't help at all if you're, for example, trying to find out what
options you can pass to a binary. And don't try to argue that some
binaries don't have options: you still need a man page to tell the user
that, so they don't have to wonder about it. At any rate, once you've
gone to the trouble of writing a useful description for README.Debian,
you're most of the way to having a man page anyway--except that you've
stuck it in a place that's not immediately obvious, can only be found if
you know which package the binary came from, and which isn't integrated
into the normal documentation system. How is that helpful?
I really don't understand this hatred of man pages.
Mike Stone
Reply to: