On Fri, Oct 20, 2000 at 11:23:54AM -0400, Chris Gray (cgray@tribsoft.com) wrote:
> On Fri, Oct 20, 2000 at 02:27:04AM -0700, kmself@ix.netcom.com wrote:
> > Better more data than less, particularly if I ask for it. I've been
> > getting mightily peeved lately by people who post "how do I solve foo"
> > to which I reply "post output of <command list>"...which never happens.
> > I'm going to start killfiling idiots like that real soon, then I'm going
> > to take out hits on them....
> >
> > But enough of that.
<...>
> When you say "post output of <command list>", you probably have some
> idea what's going on but want confirmation (otherwise, you'd just be
> adding noise).
>
> So instead of just saying "post output of <command list>", say "if
> <command list> returns foo then do bar, if it returns baz then do quux,
> otherwise post it and let me have a look at it." The games of 20
> questions and gratuitous posts of outputs do not make for interesting
> reading or even very helpful help to the people that asked in the first
> place.
Extreme disagreement.
If the person had enough experience (to be completely distinguished from
intelligence, these are totally different aspects [1]) to know what to post,
they would have done so in the first place. As he or she did not, I'm
asking for the data I would look for on my own system to resolve the
same issue.
The simple truth is that email support puts the supporter and the
supportee at two ends of a high latency, and often low-bandwidth, pipe.
I can't see the problem they have, they can't see the range of solutions
(or hints -- I'm frequently wrong, but elimination of possibilities is
part of the process) I may possess.
The best exposition of this problem I've seen in recent years was by
Jeff Covey at Freshmeat.net on how to report a bug, and issues people
should be aware of:
http://freshmeat.net/news/2000/02/26/951627540.html
So: I ask for diagnostic outputs which may point to the problem at
hand. Frequently I find that it's some basic fundamental step which has
been missed, and which has been completely overlooked by the original
poster. Diagnostics help. If the output volume is too high, I *will*
trim extraneous data. Better that the original be posted (and eyes
other than mine see it), *then* have editing by someone who (maybe)
knows what he's talking about, than have random redaction by a newbie.
> This is just a suggestion and you in particular do a very good job in
> helping a lot of people so I hope you can take it in the spirit that it
> is offered. And of course it is a very general suggestion that I know
> won't work in all cases.
Thank you.
> Finally, killfiling lurkers is going to make for a large .procmailrc but
> is not going to do all that much.
It's not the lurkers, it's the idiots who don't realize they've got to
help me help them. I can't see their problem or state: I need more
data. And yes, I usually give a second chance -- though I range back
and forth through levels of irritability. Third requests are usually
last ones. My time is freely offered, not reimbursed. I prefer not
wasting it.
...and if it wasn't mind-bogglingly obvious, taking out hits on idiots
was a bit of hyperbole (though I have to say the thought warms my heart
in dark moments....).
My philosophy on support is to provide hints and guidance, not a crutch.
If the appropriate response is to read the man page on some command,
I'll say just that: "man foo". I *don't* believe in flaming someone on
their first request for assistance unless it's truly mindless -- there
was a thread this week in the OpenBSD misc mailing list launched by Theo
de Raat, prominently featuring the phrase "too low a lifeform"[2].
That's completely uncalled for, IMO. Though I may chide, I try to avoid
direct insult (unless it's richly deserved). Even where I disagree, I
try to do it as here: state disagreement, support assertion.
I'll occasionally write a longer reply when it's philosophy more than
mechanics at hand (e.g.: here). Some of these I've compiled into FAQs
and micro-HOWTOs, currently I've got prepared texts on backups,
GNU/Linux books, partitioning, XDM disabling, SSH RSA key authentication,
Samba mounts, and sudo. Questions I've answered enough times, in enough
depth, to have a standard response. Jeff's Bug rant is close to
becoming same.
--------------------
[1] And yes, as a matter a fact, I *do* remember wondering what all
those weird "/" things were that I had to type, once upon a time, many,
many moons ago, when the world and I were young.....Ignorance is
curable. Stupidity isn't. I'm willing to assume the former until
proven otherwise.
[2] Theo's post:
http://www.sigmasoft.com/~openbsd/archive/openbsd-misc/200010/msg01040.html
--
Karsten M. Self <kmself@ix.netcom.com> http://www.netcom.com/~kmself
Evangelist, Opensales, Inc. http://www.opensales.org
What part of "Gestalt" don't you understand? There is no K5 cabal
http://gestalt-system.sourceforge.net/ http://www.kuro5hin.org
GPG fingerprint: F932 8B25 5FDD 2528 D595 DC61 3847 889F 55F2 B9B0
Attachment:
pgpYeKIGbt59q.pgp
Description: PGP signature