
	Weekly chat #37: Tuesday 7th November 2000
	
Participants: Peter Kretschmar, Niels Joergen Westergaard, Carol Anne Oxborrow,
	Allan Hornstrup, Stefan Larsson, Sami Maisala
	
Subjects discussed: Interactive Background Analysis
		Binning standard
		Using OBT flag in STATUS bit
		Error codes in jemx.h
		Obsolete data structures
		CVS repository
		V2 deliveries and updates


Action Items: None


**** Logging Started : Tue Nov 07 09:31:01 MET 2000
> Hello chaps! Ready to chat?? END
<njw> Yep END
> Why is it that the people who have an hour extra to get to work are always
> the last to log on? Just a socio-temporal question. END
<njw> They are probably already deep into coding or ... END
*** larsson (~larsson@tenma.dsri.dk) has joined channel #jemxadr
> Hello Stefan: 5 down, 1 to go. END
<larsson> Morning everyone. END
<njw> Good morning, Stefan END
> could we save a little time by beginning with the AIs on people
> who are here? So how about NJW telling us what he's completed on
<peter> Fine with me. END
> the consolidated AI list while we wait for Sami? END
> NJW? END
<njw> Just a second - END
*** Allan has left channel #jemxadr : (Allan)
<njw> We have prepared a small introductory note for the needs/requirements
*** Allan (~allan@apollo.dsri.dk) has joined channel #jemxadr
<njw> for the interactive background analysis (CAO+AH+SB+NJW). It will be circulated
<njw> lated today.
> Yes, that closes 001031_5 end
<njw> I have not done anything for the  calibration DS etc. yet END
> No, we're meeting to clear those AI's tomorrow, NJW END
<njw> Yes, its on my calendar END
> Okay Peter, what have you done on the AI list? END
> I've tried to phone Sami, and he's not in so we'll proceed without
> him END
<peter> Just a moment ...
<peter> AI16.1: gti was definitely renamed to gti_create. On lc_pick vs lc_find, no real decision has been taken.
<peter> AI16.3: Stefan and me agreed on a standard (user parameter or start of ScW), as explained in emails.
<peter> AI16.5 Nothing really done yet.
<peter> Dito for AI 16.8 (sorry)
> So what should I put in the ADD Figure: lc_pick or lc_find? END
<peter> REVII_01 On the stack of things-to-do ...
<peter> END
<peter> Call it lc_pick ... but that may change - sorry! END
> Okay, will do END
*** smaisala (^smaisala@mizar.astro.helsinki.fi) has joined channel #jemxadr
> What about you Stefan, any action on the AI list? END
<smaisala> Hi there, Sorry I'm late  END
> Hi Sami! We're just going through the AI list. END
<larsson> A cold has kept me out of business the last week, so not much done.
<larsson> 001031_3        ALL     8/11/00 comments on jemx.h DS I did last week 
<larsson> but thats about it.
> Remember 001031_6 has to be done by Wednesday, so that we're all
<larsson> I'll try to do my homework regarding background untill thursday. END
> ready for the interactive background chat. END
> Okay , Sami, any action on the AI list from you? END
<larsson> I'll do that CA. END
> Thanks, Stefan. END
<smaisala> 16.4 is under work. 
> Yes, All the AIs to do with the determination of the interactive
> background analysis should be complete after the chat on Thursday. END
> I'm finishing off AI16.6 right now: there's not as many DSs that can
> benefit from joint plotting as I had thought, especially since
> j_verify will check all the raw data DSs numerically/statistically,
> so no need for visualisation right there. END
> AI16.2 is closed: I will add a `bad OBT' flag to STATUS and add the
> `wrong-flag' functionality to the correction executables: only time
> and exeperince will show if this is useful. END
> Availibility of slew-data is hard to discuss because all the
> hardware people are running around like headless chickens at the
> moment. All the calibration stuff will be sorted out between me and
> NJW  tomorrow. END
> Last of all we have Allan, who sadly is no longer the big cheese
> in SDAST (pending Thierry's approval): do you still do AI's Allan? END
<Allan> y
> So which ones have you done? END
<Allan> Sorry, just talking to Thierry
> Okay, we'll continue with the next agenda item: error codes.
> What do people prefer: and error code start for each executable
> or error codes based on the pipeline/level - I hope you've all read
> the emails Peter and I have sent on this subject. I prefere the
> former because then wherever an executable gets moved around to in
> the architecture, the code doesn't have to go making changes in
> their code, I can just rearrange executables in jemx.h and give them
> different values. What do the rest of you think? END
<peter> I actually support this idea. END
<smaisala> Sounds like a good idea. END
<njw> Goes for me too END
> Okay, I'll go ahead an re-do the codes in jemx.h. What sort of
> standard system for constant names do you want to use Peter - I
> agree that it should be standardised. What about the executable
> name with _ERR at the end which would keep things clear but short. END
<peter> The question goes to all of us, but the proposal looks fine to me. END
<larsson> Fine. END
> Okay so that closes AI16.12 END
> Of course, it must be understood that each executable can only use
> a fixed number of error codes so that no one overruns into other
> peoples codes. How many different codes does each executable need?
> END
<peter> 100 should suffice but make it 1000 so they are nicely apart. END
<njw> Until now I my maximum number is around 10, so allocating 1000 will be more
* njw than enough END
> We don't have room for 1000 per EXECUTABLE, originally it was
> 1000 per pipeline and 100 for each executable. END
<njw> 100 per executable is also enough END
<peter> Why not? We have the error space -300,000 to -399,999. Even if we leave room for scripts this should allow 1000 per
<peter> entity.
<peter> END
> Let's go with 200 per executable so that everyone starts on an
> even hundred value. END
<peter> I like that less - I would prefer to have a unique 'relevant digit' encoding the executable or script in
<peter> question END
<peter> Sorry, that should be two digits ... END
<peter> So either we keep 100 (which should be ample room) or 1000 per entity (executabel, script, spplication). END
> Looking at the old JEMX.h  each executable had 100 error codes,
<peter> So what does the rest think? END
<larsson> Lets go for 100 then. END
> and I think this should be enough - we never ran up against an
> error code shortage. Then that leaves lots of space for script and
> application errors. So 100 it is END
<peter> OK, then let's keep 100 per executable and use the 1000 and 10000 digits as encoding of analysis step. 
<peter> The scripts would then use the ....9xx codes always. END
> That sounds good. END
> Next on the agenda: Data structures: I hope you've all taken a good
> look through the jemx.h listing and the data dictionary I sent you.
> Is anything missing? Is anything included that is obsolete? END
> For example, can anyone tell me why -POSI-MOD  -POSI-CAL
>   -SPAG-MOD  -SPAG-CAL
> are not included in RP2.0 ? I'm just checking up here to make sure
> that nothing's missing that should be in there. END
<smaisala> My data structure names are changed. I'll send all changes to you after chat. END
<njw> Please remind what -POSI-MOD was supposed to contain END
> That would be great sami END
> I think it was the old name for the position corrections, but it
> could have been something that someone else was using. END
> Okay if there are no other comments on data structures, I'll assume
<njw> Probably replaced by -CORX-MOD and -CORY-MOD then. I don't use -POSI-MOD. END
> that jemx.h and the data dictionary are up to date for the time being END
> News on the CVS repository: I'm waiting for Allan to set up the
> server so that you can login and download/upload files as you wish.
> I'll let you know when the repository is up and running. Any
> comments? END
> Next on the agenda: V2 updates and deliveries. Sami made some
> update deliveries just yesterday, hows things going NJW and
> Stefan? END
<larsson> Since I lost the last week. My statement of "end of week for
<njw> j_ima_cor_intensity is very close to delivery. Actually I just need some
<larsson> j_src_lc and another 2 weeks for j_src_spectrum is still valid. END
<njw> test data and for that I needed to change j_ima_src_find which is also about
<njw> ready.
<njw> Sideremark to CAO: hence the question about PI channels starting counting
> Anything to be said on V3 coding and ADD5.1? The latter's more or
* njw with 0 (zero) END
> less a dormant subject until the CVS repository's in place? 
> Comments? END
<njw> N
<peter> No
<smaisala> No
<larsson> No
> Any other business? END
<larsson> No. END
<njw> N
<peter> Nothing I can think of at the moment END
<peter> I'd just like to point out that I need those missing executables ASAP for integration to continue. END
<peter> But you knew that, I guess ... END
> Yes we did peter. END
<njw> I knew that, and I heard you END
> Okay then, thanks for a nice brief meeting. Have a good and
> productive week everyone END
*** njw has left channel #jemxadr : (bye bye)
<peter> Thanks, 'see' you soon again. END
<smaisala> BYE
*** Signoff: larsson ()
