Report bugs to IBM

[Christian Motz]

In article <6136 at hub.ucsb.edu> 3003jalp at ucsbuxa.ucsb.edu (Applied Magnetics) writes:
>In article <3197 at awdprime.UUCP> jeffe at sandino.austin.ibm.com (Peter Jeffe 512.823.4091) writes:
>
>>  [ Report your bugs to IBM, they'll get fixed. ]
>
>OK, I'll report them through channels...  but the many layers of
>bureaucracy within IBM can oppose stupendous inertial forces.  The
>parallel path through Usenet is irreplaceable.

Hear hear! Reporting a bug to IBM and getting it fixed by them can  be
a very frustrating,  tiresome,  not  to  mention  lengthy  process.  I
might illustrate that using one of the problems I  had  at  work  some
time ago.
   Every time I used Distributed Services message queues on our RT PC,
things would break after a certain amount of time,  i. e.  the  system
would come down with a panic and a kernel dump.
   No problem, I thought, I'll  just  call  my  friendly  IBM  support
people, and have them open up a problem record for  me.  This  I  did,
and, knowing that they would be needing it, sent the  kernel  dump  to
them even before they specifically requested it, thus saving some time
in the process.
   Then the fun started. They called back, basically to ask  me " what
did you do to provoke this" -- which, of course, is the wrong question
to ask. After all, it isn't *MY* fault that the kernel  dumps,  or  is
it?. It took some time  to  explain  to  them  that  the  machine  was
crashing completely, with only the panic message displayed.
   Five minutes later, they started  again. " What  kind  of  messages
indicate that you have a problem?" was the question. At that  point  I
was close to the boiling point -- after all, I had just  explained  to
them what had happened, and if they would have the slightest idea what
UN*X is all about, they would have gotten the point.
   A couple of phone calls later, they finally seemed to have an  idea
what I was talking about -- but now they acted as  if  they  couldn' t
believe there was a problem. They obviously wanted to get it confirmed
by someone of their own company. Thus, they sent me two SE's that  did
nothing else but watch while I reproduced the problem. Only  then  did
they believe that the problem really existed.
   Meanwhile, they had received the tape with  the  kernel  dump ( and
were very surprised that I had sent that on my  own  initiative),  but
they needed more (of course --  I  had  expected  that,  and  there' s
nothing wrong with it). Ok, I said, give me a modem number, I'll  send
it to you. They almost said "what's a modem?" ;-) --  naturally,  they
didn't have one.
   Ok, I said,  if  you  could  find  me  someone  at  Austin  with  a
Trailblazer, I'll send it to them directly (since IBM Germany would do
this anyway with all the stuff they got), regardless of the  DM  50,--
or so in transatlantic phone charges this would result in. "Uhm,  I' m
not sure we can do this ..." was the reply.
   All in all, this cost two days -- and remember, this was  only  the
fight with IBM Germany. I do not doubt that IBM has some great  people
at AWD in Austin and Palo Alto; as a matter of fact, I know  it  since
we have them here on the net. But as long as IBM  does  not  have  the
people for proper UN*X support, we need the contact to the development
people, and we need it *BAD*.

>Lest this be perceived as a flame, here is the talisman:  >>  :-)   <<
>Whatever happens at IBM Austin, STAY ON THE NET!  We need you.

Amen! Ooops -- forgot for a second that I am an atheist ... ;-)


Overheard at an AIX V 3 Installation Workshop:

Question from Participant: Will IBM Germany hook up to USENET sometime
soon? Answer from IBM Representative: What's "USENET" ... ?

--
Christian Motz                                root at dialog.stgt.sub.org