Re: goddammit
On Fri, Feb 02, 2007 at 05:19:12PM +0100, Wim De Smet wrote:
> On 2/2/07, hendrik@topoi.pooq.com <hendrik@topoi.pooq.com> wrote:
> >On Fri, Feb 02, 2007 at 01:59:30PM +0100, Wim De Smet wrote:
> >> On 2/2/07, Incoming <s.clement@cashette.com> wrote:
> >> >Before anyone decides to engage in infantile gibberish, otherwise known
> >as
> >> >rants or flames, don't waste my time.
> >> >
> >>
> >> Oh the irony!
> >>
> >> There are nicer ways to ask for help you know. Starting of by flaming
> >> people, then preemptively trying to make them refrain from flaming you
> >> isn't going to help much.
> >
> >I didn't read any of it as a flame -- more like a scream of agony.
>
> I read "this thing sucks, you need to be a systems engineer to be able
> to use it". Sounds like a flame to me. But no matter.
More like vividly expressed frustration. After all, he didn't say
anyone was an idiot, or accuse anyone of using Spongebob to soak up
left-over baby's blood, did he?
> >
> >The trouble with Debian is that is is an old distro, and some of its
> >documentation has become seriously out-of-date. If there was some way
> >of maintaining the documentation along with the code, that would be
> >great!
Young distros haven't had time to let their docs get out of date -- they
either don't have any or have wrong docs from other distros.
>
> I second that. The docs on the site are somewhat out of date. The
> stuff in /usr/share/doc is usually up to date though, so is the
> install guide.
I think there's an interesting research project here -- how to keep
documentation up-to-date in a volunteer organisation. Or in any
organisation for that matter. Developers seem unable or disinclined to
do so. This leads to documentation specialists, but unless they also
develop, they won't have the information they need to write into the
docs. In commercial software houses, where a manual is required, one
ends up with neat documentation that is of very little value.
Now the Linux documentation has escaped most of this, but it is not
complete, and not up-to-date.
Time to brainstorm, I suspect. Any ideas (like dependency tracking,
maybe) how to at least identify what documentation is outdated? If we
can't do at lest that, we're doomed.
-- hendrik
Reply to: