[evla-sw-discuss] [Fwd: Re: Engineering Software Requirements]

[Bill Sahr]

A bit more on EVLA Engineering Software Requirements.

Bill

-------- Original Message --------
Subject: Re: Engineering Software Requirements
Date: Tue, 11 Mar 2003 10:50:12 -0700
From: Robert Hayward <rhayward at aoc.nrao.edu>
Organization: NRAO
To: bsahr at cv3.cv.nrao.edu
References: <3E6E15E8.254C272F at nrao.edu>

Hey Bill, fine by me.

One other comment I might make, based on our experience with the VLBA 
CHECKER system (which I'm assuming is already superior to the VLA 
version) is the difficulty in changing the error limits. For example, 
the CHECKER routine is suppose to generate a flag if an LNA bias voltage 
exceeds +10 millivolts. This was fine for the old GaAs amps which always 
had monitored biases less than -100 mV. Our newer InP amplifiers 
routinely have biases that are slightly positive, so an upper error 
limit of +100 mV would make more sense. However, if my understanding is 
correct, the limits for the error checking seem to be embedded right in 
the software, rather than in an external look-up-table. This means the 
entire program has to be recompiled to invoke this type of change, so 
it's a pain to do and doesn't often get done "on-demand". So we 
constantly get maintenance forms written up against these erroneous faults.

So just a plug for a much more user-friendly and flexible CHECKER system 
for the EVLA. Of course, if the MIB is doing all the error checking, 
then it might be just as painful to update its flagging procedures.

Cheers,
Bob


Bill Sahr wrote:
> Bob,
> 
> I thought your comments for the EVLA Engineering 
> Software Requirements were very good.  I hope you
> don't mind that I decided to post them to one
> of the EVLA software discussion exploders.  I 
> did so because I want the people who are thinking
> about and working on MIB software to have access
> to your comments now rather than in a few weeks
> from now.
> 
> Bill Sahr
> 
>