From denault@jeans.ifa.hawaii.edu Wed Aug 22 11:38:43 2001
Date: Wed, 22 Aug 2001 11:38:20 -1000 (HST)
From: Tony Denault <denault@jeans.ifa.hawaii.edu>
To: John Rayner <rayner@jeans.ifa.hawaii.edu>
Cc: Tony Denault <denault@jeans.ifa.hawaii.edu>
Subject: spex software issues summary

Hi John,

I'm not sure I fixed anything, but I wrote up some comment on my spex
work:


 1. GOs Timeout.

   Was asked to look at spex/DSP problems.

   For some reason, the DSP doesn't complete its clocking/acquistion during
   the required time frame. It is unclear what the causes is, but delays of
   over 1 second are encountered. Eventually the DSP does complete and acquire
   all the pixels, however, the IC has already declared a timeout.

   I've relaxed the timeout constraints somewhat on the IC...so maybe
   it may work better but extremely long delays will still timeout.

   I am concerned that the causes of these delay are not understood.

   Re-compile and installed version 2001.08 on 8/22/2001.

 2. XUI Freezes

   There is a 1 minute freeze, John experience once. And another
   freeze John encounters more frequently which requires restarting
   the XUI to fix. During all my testing in hilo, I could not reproduce this
   problem, although John encountered it frequently when running from his
   office.

   I have no comment on this problem.  It is still a mystery.

 3. spextool

    IDL is installed and configured correctly on all summit workstations.
    Spextool is install on all spex workstation.
    I tested running spextool as a guest users account on all
    summit workstations: stefan, boltzmann, wien, kelvin, and helmholtz.

    After email from Mike C., I looked at the file permissions on all summit
    workstations. Write for a path.txt file in Spextool/data needed to
    be changed on boltzmann, wien, kelvin, and helmholtz.

    I suggested to Mike that a user perference file should be written to
    home directories, and having common file in the system area is the
    wrong approach.

    As for the trouble reported in Howell's feedback form (8/16),
    I believe IDL and spextool was working properly during those
    data. And it could have been operator error.

  4. tony's ranting: system non-maintenance.

   It always takes some polishing to get a system free of minor
   operational glitches.  This has been true with all systems, and it
   should be standard practice when dealing with these types of complex
   systems.  Unfortunately for Spex, its break-in time was cut short.
   It's like many IRTF telescope-installed systems, the IRTF doesn't
   have the staff to properly care for and maintain the systems.  As
   part of on-going maintenance, I believe time should be spend keeping
   system as current as possible. Part of guidedog's CPU upgrade was
   to update the software layers used by the application. These software
   layers are the OS (2.6 to 2.7), Ixthos development tools (4.2 to 5.2),
   UNIX compilers and development libraries. I was hoping to upgrade these
   software layers on both bigdog and guidedog.  I only have direct control
   over the source code I write, but use many 3rd party components
   that contain bugs, so it is impossible to guarantee a perfect system.
   Using the best components available is a continuous job.
   Although this doesn't address specific concerns, I believe it
   does produce a higher quality systems.

Tony


/-------------------------------------------------------------------------\
| Tony Denault                       | Email: denault@irtf.ifa.hawaii.edu |
| NASA IRTF, Institute for Astronomy |              Phone: (808) 933-0893 |
| 640 North Aohoku Place             |                Fax: (808) 933-0737 |
| Hilo, Hawaii 96720                 |                                    |
\-------------------------------------------------------------------------/