[daip] FILLM CPARM(7) default change request
Michael Rupen
mrupen at aoc.nrao.edu
Wed Jun 27 13:48:34 EDT 2001
Dear DAIP,
the FILLM default for CPARM(7) [FQID assignment] is unfortunate for most
spectral line data. Currently CPARM(7)=0 gives one FQID for *all* data,
while CPARM(7)=-1 assigns a separate FQID to those data which differe from
one another by more than the maximum offset expected due to Doppler shifting
between sources 180degs. apart on the sky. Most VLA observers in my
experience want CPARM(7)=-1, and neophytes often wind up having to re-FILLM
their data after using the defaults. This is not a big deal (RTFM ;) except
in cases where the observer does not have access to the original data tape,
because (s)he has snagged the data via FTP. This is happening much more
frequently these days, as bandwidths improve, and the analysts are now
loading even spectral line experiments for internet transfer (Meri tells me
they get about one spectral line request per day). They tend to use the use
the defaults in FILLM unless explicitly requested to do otherwise; when new
users start working with those single-FQID data they can get quite confused.
Of course I've asked the analysts to use CPARM(7)=-1 from now on, but
in any case it would seem logical for the default to be the action most
users would most often use :)
So, I propose switching the behavior of CPARM(7) for spectral line data,
so that
CPARM(7)=0 --> assign new FQIDs if the freq. changes by more than the
expected Doppler shift for widely-separated sources
CPARM(7)<0 --> one FQID for all data
CPARM(7)>0 --> same as now
The main argument *against* this is that current SuperPundits will be
surprised by the new behavior. This doesn't seem a very good argument,
since one would rather have confused SuperPundits than confused grad
students.
Of course there was probably a good reason for the initial choice of
defaults, so do let me know if you feel this is not a good idea!
Cheerio,
Michael
More information about the Daip
mailing list