From ggriffin@cmbr.phys.cmu.edu Mon Aug 10 18:17:59 1998 To: Giles NovakCc: Bob Loewenstein Subject: talked to bob Giles, Bob and I just talked. You and he may already have talked about this, but I just thought I'd summarize the plan that is taking shape, to sort it out in my own head, 1) control pc (taking it's orders form a tcl script) tells comsoft to move to the start of the scan comsoft<->control sparo 2) control pc tells sparo mac to start, also sends sparo info on the current location, the kind of integration it wants, etc. control->sparo 3) sparo tells control, "nod the telescope left" control<-sparo 4) control takes care of nodding the telescope (issues comsoft commands, then polls till it gets there) comsoft<->control sparo 5) control tells sparo, "ok, we're there" Also tells sparo, "that was a left nod" just so there is no confusion. control->sparo 6) sparo does its thing, then says to control, "nod telescope right" control<-sparo 7) again control takes care of it, then comsoft<->control sparo 8) control tells sparo "ok, we're there" control->sparo 9) sparo does it's thing, then says, "done" control<-sparo Well, this is all easier said than implemented, but at least there are only 4 sorts of communication: 2) control telling sparo to start + sending info to sparo 3,6) sparo telling control to nod 5,8) control telling sparo, "we're done nodding", and confirming that it was a left or right nod 9) sparo says, I'm done. Also note that sparo need never send commands to, or poll comsoft: the control pc takes care of that. When I say, "nod left", or "nod right", that is really handled by the "movb1" and "movb2" commands in comsoft, which move between two pre-defined beam offsets. Call all of the above an "integration". The control pc's tcl script would have to know where you want it to move for each integration. Then, your software assembles all these "integrations" into an image? -Greg