From brunswig@iram.es Fri Nov  9 14:00:17 2001
Date: Fri, 9 Nov 2001 13:52:48 +0100 (CET)
From: Walter Brunswig <brunswig@iram.es>
To: Hans Ungerechts <ungerechts@iram.es>
Cc: Albrecht Sievers <sievers@iram.es>
Subject: Re: NCS & CCS: Online Data Processing

On Fri, 9 Nov 2001, Hans Ungerechts wrote:

>
> some comments / questions about W's project (2001-11-07):
>
> 2.2:
> ... shall send data ...
>
> obsProcedure:
> how do you send the parameters of the obs procedure to the
> processing software?
it has to be checked if we have to transfer also parameters.
The requested information might be already in the header. If not,
we can put a few more words into the header. If more is needed then
we have to think about it. (see also design/obsProcedure in v.2001-11-09)
>
> 4.3:
> do you indeed keep otfcal running? how? is this safe?
> what do you do if otfcal crashes, hangs up, gets confused?
right now we restart otfcal when the project changes. my script
- could detect if otfcal crashes, inform the observer and restart it.
- would hang if otfcal hangs
- does not know what to do if otfcal "gets confused"
in the future, the whole scheme shall also be flexible enough to
allow a frequent project changes with queue operation.
>
> i would prefer "scripts" instead of macros
ok
>
> instead of "currently developed" say: "are under development"
> or "are being developed"
ok
>
> general questions about the scripts for otfcal:
>
> -- should they handle subsans or completed scans?
for most observing modes both cases have to handled
> -- can they use intermedate files for the data?
yes, we woud have to define some rules (like specific work directories)
as several postprocessing task might run at the same time. i add a
requirement.
> -- coding standards?
to be done. I will look into it (with help from you or Albrecht?).
> -- recovery from problems?
at least, things should be such that offline data processing could be
possible as well. the operator should have the possiblity to restart
the post processing software
> -- consistency/ sanity checks?
any suggestion ?
> -- for each observing mode:
>    what should the script do?
has to worked out of course
>
> Hans
>

Your suggestions have been worked into v 2001-11-09 of the doc, any
further points we shall discuss in the meeting on monday. Are other
astronomers interested in this topic ?

Walter

Walter Brunswig, IRAM,  Avenida Divina Pastora 7-NC,  E-18012 Granada, SPAIN
e-mail: brunswig@iram.es    phone: (34)-958-228899	fax: (34)-958-222363