From novak Fri Jul 10 13:04:02 1998
To: cdd@socrates.patnet.caltech.edu
Subject: Re: Yerk

Darren,

Thanks for your reply and offer to help get Vax code.  

I have a few 'general' questions about the nature of the communications 
between Yerk and Vax programs.

(1) 'Polling' vs. 'ask and wait':

Concerning the interaction that takes place when the Mac requests a 
telescope move: Does the Mac send a command and then keep 'polling' the Vax 
until it gets there, or does the Mac 'move' command get followed by a 
period of time during which the Mac waits for the Vax to send a command 
back saying "I've arrived".  Bob L.  says the KAO code worked in the 
"polling" method, but I don't know if Steve (or you?) changed this to
conform to CSO scheme.

(2) Mac 'always' initiates an interaction between Mac and VAX.  ...right?

You said that it is the Vax that is listening for commands to move.  Is 
there ever an unsolicited communication sent from the Vax to the Mac (i.e.  
"ERROR: I'm no longer pointed at the source") ?  Or would any such error 
only be reported to the Mac when the next command from the Mac arrives?

(3) No 'checking', right?

Does the Mac ever 'check' to see that the telescope is still on target?  Or 
does it just assume that once the telescope has successfully completed the 
move (and reports this to the Mac) then it will successfully remain pointed 
at that postion until commanded to move again ?

Bob said that the KAO program did no checking.  It did not learn about any 
loss of track that occured during an integration unitl the end of the 
integration when it attempted to move again and did not meet with success.

(4) Vax does absolute moves, Mac does relative moves.  ...right?

My recollection is that the command to move to a new source was entered 
into the VAX.  But the Mac issues the commands to move to 'right beam' and 
'left beam', and also to update the pointing corrections.  Is this the 
general idea: you input into the Vax the target coordinates and then the 
Mac tells the Vax to modify the 'offsets' to these coordinates?

If this is correct then I assume that the only way the Mac knows what
source we are looking at (i.e. what exact coordinates) is that it
requests that information at the beginning of a file and puts it
in the header.  Is this correct?

(5) Does the beamswitch command look like this:     'right beam now'
or like this:                    'go over to az-offset 789.2 arcsec'

Thanks again for your willingness to help me out.

Giles


From cdd@socrates.patnet.caltech.edu Fri Jul 10 17:08:27 1998
To: novak@clark.phys.nwu.edu
Subject: Re: Yerk


Giles,

Your guesses are mostly correct:

(1) 'Polling' vs. 'ask and wait':

Polling is what we do at the CSO.  All commands to the telescope receive in
principle an immediate response.  To check that the telescope is acquired, a
separate command is sent (repeatedly -- hence polling).

(2) Mac 'always' initiates an interaction between Mac and VAX.  ...right?

Yes.  For Hertz, there is a not an error connection initiated by the VAX.
SHARC does have such a connection, though.  For active observing at the
CSO, we don't really need the errors sent to the Mac since they pop up on
the VAX display, but for something like your remote observing at the Pole,
that might be a good idea.  Either in the form of frequent error status
requests from the Mac to the Telescope, or a link where the Telescope sends
error signals to the Mac unsolicited.

(3) No 'checking', right?

Not a lot of checking.  I think the only way data acquisition would stop at
the CSO is a lost TCP/IP connection, telescope never entering acquired state,
or a garbage response to a TCP command.

(4) Vax does absolute moves, Mac does relative moves.  ...right?

Maybe I wouldn't divide it that way, but the positioning of the telescope is
done half by the VAX and half by the Mac.  Do to pre-existing capability,
poiting at sources in a catalog, precession, focusing, and the possibility
of some fixed offsets (like RA offset) are done by the VAX.  Things specific
to Hertz are done on the Mac, like beam switching and figuring out pointing
corrections for different pixels and different instrument rotations.  All of
the positioning commands are absolute (although they may be offset commands).
For example, we say "Change the azimuth offset to +200 arcsec" and not "Change
the aximuth offset by an additional 10 arcsec".

(5) Does the beamswitch command look like this:     'right beam now'
or like this:                    'go over to az-offset 789.2 arcsec'

The command is 'go over to az-offset 789.2 arcsec'

Darren