
     SDAST CHAT MEETING #63: 29th August 2002
     
 Participants: Carol Anne Oxborrow, Peter Kretschmar, Niels Joergen Westergaard,
 		Stefan Larsson, Sami Maisala
 		
 Action Items:
 
 AI020828_1	ALL	6/9/02 Email NJW with list of dates when one is
 OPEN			available/unavailable for duty at ISDC
 
 

 Subjects discussed: SDAST meeting November 28+29 2002 at ISDC
			Operations training at ISDC 16+17 September
			Rotation schedule at ISDC
			Instrument representative at ISDC during Christmas
			Outstanding SPRs: j_src_spectra, j_ima_src_find
			j_bkg_scal_calc, j_ima_cor_intensity, j_cor_position,
			j_dead_time_calc
			Limits of time correlation determination with DAL3GEN
			Detailed component descriptions for Peter's presentation
			next Wednesday, and for Masha's UM
			AI list: lots of old AIs closed
			Jerome Chenevez: Mr. NEW
			j_ical: when to deliver
			Interactive/offline tools in general


*** njw (~njw@130.226.216.2) has joined channel #jemxadr
*** peter (~pkretsch@isdcsf3.unige.ch) has joined channel #jemxadr
*** larsson (~larsson@tenma.dsri.dk) has joined channel #jemxadr
<peter> Hi everybody! END
<larsson> Hi Peter et.al. END
> Hello boys - it's good to see you both so punctual as usual END
> Hi Stefan - it's good to see you too! Now the big question? To wait or not
> to wait? END
*** smaisala (~smaisala@gemini.astro.helsinki.fi) has joined channel #jemxadr
<smaisala> Hi all. END
<peter> Today it goes really well! Though Silvia obviously has not taken a 
laptop to the beach :-) END
> So, no waiting today! Just one warning: our system's been locking up suddenly
> the last couple of weeks - it usually last only 5 or so minutes, but it can
> mean we at DSRI are suddenly cut off - but I'll rely on Peter to keep things
> moving if that should happen. END
> First point on the agenda: get your calenders out for the next SDAST meeting
> in ISDC: does November 28+29 suit everyone? END
<njw> OK for me END
<larsson> Yes madam. END
<smaisala> Fine with me. END
> Peter? You will presumably be present? END
<peter> I have little other choice :-) END
<njw> Then it is settled, I believe END
> Okay so point 2 on agenda: Operations training at ISDC. NJW and I will be 
there
> 16+17 to be trained and take part in the JEM-X simulation: what about the 
rest
> of you? END
<larsson> Teaching and other "pleasures" on my agenda. END
<peter> Well, I will be absent, which prompted me originally to invite you. END
> You invited us knowing you'd be away at the time - If I was more sensitive
* oxborrow person, I'd be quite hurt by that statement. END
> And Sami's without funds for travelling. Okay NJW and I will pay particular
<peter> The idea was to have JEM-X specialists around in my absence. And ISDC 
and you can train how to deal with JEM-X without PK. END
> attention so we can share the training with the rest of you. END
<peter> Pascal will be there and act as my 'second'. END
<peter> But finally, as Nami wrote, it is usually only moderatly exciting. END
> Okay, point 3: Rotation schedule at ISDC: Mr NEW is now called Jerome, and he
> will share my office at DSRI. Any comments or problems with the rotation? END
<njw> The rotation schedule seems to be agreed to by all the present honoured
<njw> members of the SDAST; however, Sren Brandt is on holiday and has not
<njw> responded yet.
<njw> The period around Xmas is not so easy to cover. But can I ask you how much
<njw> should we plan now for january/february and march ? Or to put it
> Do you mean actual Christmas, or just either side of it? END
<njw> differently: how much time in advance would you like to know the schedule? END
<njw> I meant Dec 22 - Jan 02 END
> If I can lure my parents to Copenhagen for the festivities, with grandchildren
> as bait, I can maybe be at ISDC over Christmas. END
<njw> That would be great if it appears that JEMX will need continuous attention END
> Okay further discussion can probably take place offline if everything really is
> in place up to Dec 22. I don't doubt that JEM-X will need continuous attention -
> it's a wonder we're not launching a person up there with her. END
<peter> Personally, I will be in Germany for my mother's 70th birthday and Xmas (same day) from Dec 23-25 but plan to be generally around from Dec 26 to 02. This does not mean though that I want to work continously during that phase. ...
<peter> END
<njw> Wait a bit, when do you expect the planning for Jan/Feb ?
<peter> I think we should have some planning soon even if we modify it later - what do you think? END
<njw> All right, I'll wait till Sren comes and then make up the long-term planning.
<njw> It means also that I'll need your input regarding unavailability periods 
END
> I'll investigate the grandparent option: the children MIGHT find it fun to
> hold Christmas on 2nd January with extra goodies from Geneva. END
<peter> Action on all to sen NJW periods of unavailability, right? Due next 
Friday? END
<larsson> Thats OK (although I will have some uncertainties there). END
<njw> It'll be uncertain anyway END
> Even with alot of scheming, my end will be pretty uncertain till October. END
> So point 4: outstanding SPRs: Peter, your comments first. END
<peter> Well, I'm waiting for quite a lot of deliveries with different urgency:
<larsson> So I am not the only one???? END
<peter> High priority - j_src_spectra (SPR 1717,1838; SCREW 747); j_ima_src_find
 (SPR 1837; SCREW 660,748)
> No Stefan, you're not the only one!!!! END
<njw> My delivery of j_ima_src_find is imminenent and will close the SPRs/SCREWs
<njw> mentioned END
<larsson> [Thank God (and CA)] END
<peter> Lower priority - j_bkg_spec_sel (SPR 1279); j_bkg_scal_calc  (SPR 1728;
 SCREW 724); j_ima_cor_intensity (SPR 1509)
<peter> Even lower: j_cor_position (SPR 1859) and j_dead_time_calc (no SPR? but 
there are these Log messages with pointer values) END
<larsson> j_src_spectra is close....as it was early in the week too... but its 
closer now. END
<peter> Should I write an SPR against j_dead_time_calc?
> Those log messages are for debugging only: they disappear with little values
> of chatty. END
> Yes, Peter, write an SPR: I get nervous and restless if there isn't a steady
<peter> Yes, but I don't want to hack chatty to a very low level there. Can you just have them only for CHATTY_MAXIMAL? END
> inflow of SPRs! END
<peter> I knew that, that's why I'm trying to oblige at least twice a week :-) END
> Will do! END
> I suppose it hardly need be said that we're all working away at our SPRs as
> hard as we can, and they'll be delivered ASAP. END
> Peter: one thing I noticed, testing j_cor_gain with all kinds of garbage 
> (following your rotten example!): I set some SPEC spectra times to 0000 in the
> four OBT columns. J_cor_gain picked these up and did the right thing, namely
> use fixed default values of gain parameters, but later in the program the
> conversion to MJD doesn't appear to find a poor time correlation for these
> OBT values, can you explain this? Is 0000 infact a valid ISDC OBT? I didn't
> have time or inclination to look into this further. END
<peter> That depends on the time correlation file used. Could you make a copy of all this available to me? 
> Thanks for the SPR Peter, it's just my size! END
<peter> Which ISDC_ENV and REP_BASE_PROD did you use? END
<peter> You're welcome :-) END
> I'll  send you a little package - in exchange for the SPR - one good deed
<peter> But please remember to run your unit test with chatter below MAXIMAL then. END
> deserves another. END
> For unit tests I always use chatter<=4. END
> Any more comments about SPRs and recent deliveries? END
> Next on the agenda: 5) Detailed component descriptions: okay guys we can't put
<peter> Not now, I have not yet tested your most recent deliveries. END
> this off much longer, though I'll admit we've made a very valiant attempt at
> sweeping this task under the carpet. Can I suggest that we all get this done
> by next Tuesday (is that in time for you Peter?), so that Peter can present
> these descriptions to whoever next week? END
<peter> Tuesday is cutting it _very_ close. I'm giving my talk on Wednesday morning ... END
> Sorry, I thought it was Thursday: alright, how about all detailed descriptions
> in to Peter by Monday evening? END
<larsson> I'll have to that tomorrow, if not I will be in trouble anyway. END
<smaisala> What should we put in our _detailed_ descriptions? END
<njw> I'll try to be quicker but won't promise too much END
> I'd be glad to get this out of the way soon anyway. For each component I'd 
<peter> What I mainly need is additional material about the methods and limitations you have in the existing code.
> though to have a diagram, and algorithm description, some design remarks and
<njw> Peter, will latex format be OK ?
> a list of detailed error output and a list of limitations. END
<peter> LaTeX is fine. END
> And xfig/eps figures? END
<peter> Sami, to take you as example. You have very nice drawings already, but what I could use is a bit more detail on how you actually select.
<peter> Also how you generated your models. END
<peter> Yes, xfig and eps would be wonderful - just my tools. END
<peter> From NJW and SL especially I could use some more figures. END
<peter> Imagine trying to explain the workings of your code to a student. END
<njw> Y
<smaisala> OK. Now I know what you want. END
> Should we also forward out documentation files to Masha for the UM? END
<peter> A good idea, though you might do this in a two stage process:
<peter> (1) unorganized material slanted for a talk to me;
<peter> (2) well organized LaTeX text with figures etc nicely included (make it a section) plus the accompanying EPS files to Masha. END
> You'll accept organised material as well presumably? END
> Any more comments on detailed component descriptions? END
<peter> No 
> Last on the agenda: AI list. For brevity, I'll just asked if any AIs have been
<larsson> No
> closed. END
> I've got the material on my desk, from NL, which will allow me to put numbers
> to our gain-glitch problem, so that AI could be closed soon, though I won't
> promise anything. Is other AI work on-going? My AI(s) about performance
> parameters and monitoring parameters have been passed over to NL, who said he'd
> talk to Nami directly about these matters, which he may have done in a preliminary
> way the last time he was at ISDC. END
<peter> One could argue that these have been closed, as NL and me together have set up 'standard' OSM displays for JEM-X. END
> `Closed' is good! END
<njw> Carol Anne, is the AI page  ~oxborrow/sdast/AIList.html kept updated because
<peter> Regarding AI23.5 Find OGIP standard names for Sami's columns COODEG and COOSTR
<peter> where did these come from again (sorry, I _should_ know). END
> yes, that's the updated page, but I can only update it according to what people
> tell me they've done. END
<njw> there is a couple of actions on me (write SPR for SKY-IMA units and produce
<njw> a rotation plane) that you can consider closed END
> I think that AI18.16 NJW/NL	ASAP Make sure IT provides a list of HK parameters
CLOSED 28/8/02			to be monitored by operators in NRT

AI19_11	NL	Aug 2002 Talk with IT and Nami about what derived HK parameters
CLOSED 28/8/02		 	should be used for performance monitoring


> Sorry, major cut'n' paste mistake: I think that COODEG AI is obsolete now since
> the design of one of his tools changed before we could even use these names. 
> Sami shall I cancel AI23.5? END
<smaisala> I think these names COODEG and COOSTR are history. We have already made the template so we can close this AI. END
> Do you mean AI22.3 NJW, and AI020822_1 and AI22.8? I'll close them now? END
> Sami, are you finished with AI23.8 which is yet another email with background
> handling design? END
<njw> No AI23.2 and AI020822_1 END
<smaisala> I think we can close this AI. Do you agree? END
> Yes, I think so Sami, what do the astronomers think? END
> Peter, NJW, Stefan are you satisfied with the background handling design or
> do we need another go around? END
> NJW, related to the two AIs of yours just closed, you have two remaining AIs
<njw> For the moment the design seems OK END
<peter> For the moment it is quite good, I think. We will surely rediscuss some matters in Novem
<peter> ber. END
> still open: one to write a complete PV phase plan and the other to make an
> ADD/UM section describing the different flux units for spectra and images. Just
> a reminder. END
> Okay, AI23.8 is officially history. END
> AIs 23.6 and 23.7 on Peter and  Sami seem quite urgent, are they closed guys? END
<peter> I guess yes, that was some time ago and we have an ATTI-DBG template. END
<smaisala> Yes. They are closed.END
> Well, that's a lot of lingering Ais out of the way. Now, AOB.
> Any other business anyone? END
<peter> Could you tell me more about Jerome 'New'?
<peter> END
> He's about as tall as NL though darker and younger. He's not danish though he
> speaks danish and has been poached from my husband's institute, the Danish
> Meteorlogical Institute. He is now the proud owner of a copy of ADD7.1 but
<njw> Jerome Chenevez will start here October 1 (though he'll try to get off the
> hadn't read it yesterday when I met him very briefly. He'll be sharing my office
> and starts work here on october 1st. END
<njw> hook earlier). END
> I hope he doesn't smoke ! END
<njw> Jerome is hired for a two year period (PRODEX money) to be extended if
<njw> possible END
> What is PRODEX? END
<peter> If you can send him here for 1-2 days still before launch that might help him for his first 'real' visit in the PV phase. END
<njw> Well, he has not yet signed the contract yet but seems eager.
<njw> PRODEX is some EU funds assigned to something like this END
> I'll talk to him about a little pre-flight experience at ISDC when he turns up
> again next. END
<peter> Thanks for all the information! END
<peter> I have two more AOB points:
<njw> Currently he is assigned to work with Carol Anne in the week where she'll
<peter> (1) Everyone should send MR inputs to Carol Anne very soon. END
<njw> make her V1 products go smoothly END
<peter> (2) FYI: I have j_bin_evts_lc ready for delivery, it's currently waiting for today's template update. 
<peter> j_bin_rate_lc is also quite far, but will probably take until tomorrow. END
<njw> Peter, if you want to test a predelivery j_ima_src_find-1.5.0 then pick it
<njw> from /home/isdc_guest/westerga/deliveries/j_ima_src_find-1.5.0.tar.gz END
<peter> Regarding Jerome, mythought was that we would have more time to explain small stuff (who is who? what is where?) in the calm before the storm. END
<peter> Thanks for that - I may try later today or tomorrow. END
> I have some AOB too:
> NJW and I have been talking about the currently non-existent j_ical tools.
> At least two of these will be in very rough form at launch, and we thought that
> it would be best for the users (instrument reps and hardware guys) to use
> them in a private, undelivered form, wherein I can make lots of changes and
> improvements as we need them, and then have the first official delivery well
> after launch. I think we all agreed that after-launch means the end of the
> year. There are several good reasons for this plan:
> 1) lack of time before launch
> 2) don't want to drown in SPRs by delivering crude version to ISDC
> 3) Lots of flexibility
> 4) Tools only used by cognoscenti, not general users, so shouldn't ever be
> released to general users as part of support-sw
> 5) hardware guys wary of ISDC tools/system to say the least!
> 6) because.
<peter> Here are my counterarguments:
> 7) j_ical doesn't make any end products
<peter> (1) Agreed, that _is_ a problem
> 8) because END
<peter> (2) j_ical_ will not need or get SPRs because it is offline and won't be under CC.
<peter> So you can redeliver at your whim. And you do not even have to supply a unit test for interactive tools. 
> You say that now, but what about later...END
<peter> (3) You are slightly more flexible that is true.
<peter> (4) You will _not_ have access to this tool in the Ops. Network if it is not formally delivered. This may not be a problem as we copy all the data out during PV phase, but that's a limitation.
<peter> Also you pssoibly deprive yourself of the archive of older versions (I was happy on occasion to be able to download my non-messed up versiion again). 
<peter> Regarding (2): only components in _pipelines_ are mandatorily under CC. 
<peter> (5) I don
> The no-CC, no-unittest thing sounds good. END
<peter> I don't see what it helps to have your tool "on the side". You run a high risk then of needing one window with one ISDC environment and another with your  personal one for yourtool. END
<peter> (7) means the same as pipeline - we simply do not have a reason to put this under CC.
<peter> Finally: I am using SPRs and SCREW also because otherwise we all (me in the first place) otherwise tend to forget agreed changes. END
> Well, we'll see how it goes Peter - just don't you or Isabelle hold your breaths
> on this one! END
<peter> What we want to hear is that you have some software which compiles and runs for you working on an ISDC machine.
<peter> If you are at that stage, you can also deliver. 
<peter> END
> Like I said: don't hold your breath! END
> One question, must all offline/interactive tools use ROOT or is this just a
>  possible tool for interactivity. Can one simply have a line pop up on the
> screen `Give lowest revolution number' `Give highest revolution number' etc.? END
<peter> Sure! You can do whatever you want.
> `Whatever you want' is good! END
> Any more AOB anyone? END
<peter> The problem is only if you write a tool relying on something like IDL or even Tcl that may not be part of our normal
<peter> software system. 
> I'm not thinking anything like that fancy! END
<peter> So if you need a special library, better make it part of your delivery, or - even better - try to use a supplied one.
<peter> Anything with a CLI is fine. END
> Thanks that's all I need to know. Any more comments from the rest of you? END
<njw> N
<peter> I would recomend writing a parameter driven tool, we can buoild frontends to that later. END
> If there's nothing more to say: have a good evening everyone! Goodbye END
<njw> Sideremark to Peter: I've just compiled and run the test at ISDC for
<njw> j_ima_src_find so formally it works.
<njw> Good bye END
<peter> Very good! END
*** njw has left channel #jemxadr : (njw)
<peter> Have a nice afternoon every one! 
*** Signoff: peter (ircII/tkirc)
