Weekly SDAST Chat Meeting #8 Friday 8th January 1998

Those participating: Allan Hornstrup, Peter Kretschmar, Juhani Huovelin,
	Carol Anne Oxborrow, Niels Joergen Westergaard.
	
Subjects discussed:
	AI: Countrate usage and extraction
	Updating SRD and URD
	prototype Calibration files
	Arranging meeting about background handling 
	Gerry Skinner and Jean In 't Zand as background consultants
	Collating background problem questions
	Programming: C vs. Fortran
	No Fortran 90 wrappers for JEMX
	Error codes
	jemx.h
	Developer's forum
	Coding guidelines

Action Items:


 	AH	15/1/99		Set up a meeting with Gerry and/or Jean
to be exclusively about background determination/handling issues. The date will
probably be Febrauary 1/2 1999.
OPEN

AI990108_2	ALL	27/1/99		Ponder the background handling/detemination
problem and send all questions (even `stupid' ones) to NJW so that a proper 
overview of the problem can be assembled before the background meeting with
Jean and/or Gerry.
OPEN

AI990108_3	PK	15/1/99		Inform SDAST of the result of discussions
with ISDC software gurus about the writing of `nice' ISSW, and the correct way
to do things.
OPEN


*** IRC session log started at Fri Jan  8 09:32:22 1999.
#jemxadr> Hello everyone: HAPPY NEW YEAR!!!
#jemxadr> Please forgive me if I'm a little slow today, I'm still feeling lousy
+.END
#jemxadr> I'll just go and rustle up Allan. END
<Peter:#jemxadr> To CAO: I'm sorry to hear that. Let's keep this brief and you 
+should ponder to keep this a short working day ... END
*** allan <~allan@apollo.dsri.dk> joined channel #jemxadr
#jemxadr> Allan should be with us soon. His watch had stopped. END
#jemxadr> Shall we get started with the Action Item List? END
<Peter:#jemxadr> Yes. END
#jemxadr> This list is available at http://www.dsri.dk/integral/doc/
<allan:#jemxadr> Should I start commenting on AI981113_3?
#jemxadr> Yes please, Allan END
<allan:#jemxadr> There is no news here. Niels has been busy with the
<allan:#jemxadr> Ballerina WS, and will probably be for some time busy
<allan:#jemxadr> with the BAllerina proposal.  Now, since
<allan:#jemxadr> the due date for that is shifted to April, I think we can
<allan:#jemxadr> get some discussions going on our subject within the next
<allan:#jemxadr> weeks.  I'll ask him and the AI is open for another couple of 
<allan:#jemxadr> weeks END
<huovelin:#jemxadr> I dont have the AI list, was it about the Background meetin
+g ? END
#jemxadr> Okay, I'll change the date on that one to 22/1/99. Is that okay with 
+you? END
<allan:#jemxadr> No this was about countrate usage and extraction. END
#jemxadr> Juhani, I'll just mail you a copy of the current AI, right now. END
<allan:#jemxadr> Next AI is about updating the monthly report. How's that going
+?
<allan:#jemxadr> END
#jemxadr> Okay, next AI is for us all forever: application-by-application activ
+ity reports each month. So remember to do it please. END
#jemxadr> AI981127_3 is a changed version of a previous AI, which now states th
+at we will each supply suggested updates to SRD and URD on a itemized basis wi
+th each and every delivery of code, rather than doing the job now and repeatin
+g or redoing th
#jemxadr> ings when the actual code is finished. END
#jemxadr> How do you feel about this? END
<Peter:#jemxadr> Seems fine to me END
<allan:#jemxadr> Fine. END
<huovelin:#jemxadr> OK with us END'
<njw:#jemxadr> Yes, it seems rational END
#jemxadr> Okay, so we leave that one open until the end of V3. Next AI is for N
+JW and has to do with providing input information in the form of FITS files fo
+r the calibration coding. I can wait a week or so with this one if you're pres
+sed NJ. END
<njw:#jemxadr> You'll get it next week END
#jemxadr> I'll update the date on that one till next week, say 15th of Jan. END
#jemxadr> Next, about the Gantt chart. Anyone find any bottlenecks, hold up etc
+, or is this AI closed? END
<njw:#jemxadr> OK END
<allan:#jemxadr> Let's close it for now.  It should be automatically reopened w
+hith every
<Peter:#jemxadr> If noone finds anything now, I'd say it's closed. From my POV 
+I found nothing. END
<allan:#jemxadr> new release. END
#jemxadr> Agreed, that's one closed then END
#jemxadr> All the other AI's have due dates well into the future, so let's disc
+uss them then. Would everyone be sure to look at the list to see if they're ex
+pected to do something that they've forgotten about? Juhani's AI on the level 
+1 and 2 data st
#jemxadr> ructures is closed I think. END
<huovelin:#jemxadr> I hope so, although the actual issue is far from being solv
+ed END
<allan:#jemxadr> right END
<allan:#jemxadr> OK ,  about a  meeting:
<allan:#jemxadr> I have communicated with Jean In 't Zand, who is not able to
<huovelin:#jemxadr> By the way, where is the latest version of the GANTT chart 
+END
<allan:#jemxadr> participate in a January meeting, but would like to join us in
<allan:#jemxadr> an email discussion or similar.
<allan:#jemxadr> I have then a confirmation from yesterday from Gerry, that
<allan:#jemxadr> he can participate in a meeting either 18-19 og 25-26 of Janua
+ry. 
<allan:#jemxadr> I have a meeting at the Ministry of Research sometime Monday a
+fternoon the 18th,
<allan:#jemxadr> so I would prefer the 25-26.  Is that at all possible to
<Peter:#jemxadr> 18-19 is (probably) bad for me - OMC meeting foreseen END
<allan:#jemxadr> you?  Further, Gerry
<huovelin:#jemxadr> I may have a problem with those dates, but I'm not yet sure
+ END
<allan:#jemxadr> would also like to participate in an internet discussion.
<allan:#jemxadr> Juhani, when can you find out if the 25-26 is a real problem?
<allan:#jemxadr> END
<njw:#jemxadr> 25-26 is fine with me END
<huovelin:#jemxadr> I'm trying all the time, and I'll let you know as soon as I
+ find out END
<allan:#jemxadr> OK. If gerry would be available Feb 1-2, would that then be mo
+re 
<allan:#jemxadr> acceptable? END
#jemxadr> 25-26th is fine with me as far as I know. In general then, are we loo
+king at an extended chat the way we did it for the ADD review? END
<huovelin:#jemxadr> It's really so unclear that I suggest you set dates, and ..
+.
<huovelin:#jemxadr> I'll try to arrange it END
<allan:#jemxadr> OK. I'll make the decision today.
<allan:#jemxadr> I suggest further, that we next week starts the 
<allan:#jemxadr> discussion on this line, including both Jean and Gerry.  That
<allan:#jemxadr> may help focusing on the relevant issues.  If they can not do
<allan:#jemxadr> internet meetings Friday, I'll let you know whne. END
#jemxadr> Okay, so AI on Allan to set up an IRC chat meeting with Gerry and Jea
+n on background issues, possibly as a Friday morning chat. END
<allan:#jemxadr> OK END
<huovelin:#jemxadr> It should be prepared well, since the approach may be so ..
+.
<huovelin:#jemxadr> different for each of us, that the comments may be difficul
+t to ...
<huovelin:#jemxadr> understand via this kind of line END
<huovelin:#jemxadr> I mean that the chats may work much better then after one .
+..
<huovelin:#jemxadr> "physical" meeting END
<allan:#jemxadr> I can see that point. We take off the AI then. END
<Peter:#jemxadr> So we try for a meeting on Jan 25/26 or Feb 1/2 with Gerry fir
+st. Right? END
<allan:#jemxadr> Right END
<huovelin:#jemxadr> OK END
#jemxadr> So we want to wait for a suitable time for an actual face-to-face mee
+ting? END
<Peter:#jemxadr> To Juhani and NJW - I wouldn't worry to get into some email di
+scussion with Jean in't Zand, even with 'stupid' questions.
<allan:#jemxadr> We are now aiming at a real SDAST meeting here at DSRI
<Peter:#jemxadr> He's a rather friendly and helpful guy. END
#jemxadr> I'll change the AI to setting up an actual meeting the 25/26th Jan or
+ 1/2 feb. END
<allan:#jemxadr> either Jan 25/6 or Feb 1/2. END
<njw:#jemxadr> Maybe somebody should write a note summing up the unclear points
<njw:#jemxadr> and 'stupid' questions so we can be focussed; this also applies 
+for
<njw:#jemxadr> a physical meeting END
#jemxadr> Who wants to collect questions (including `stupid' ones) about the ba
+ckground, I think it should be Juhani or NJW. END
<njw:#jemxadr> OK, I/
<huovelin:#jemxadr> Perhaps both of us ? END
<njw:#jemxadr> Again: I'll take the action do prepare this with the help of Juh
+ani (?) END
<allan:#jemxadr> Thanks END
<Peter:#jemxadr> Very good. END
#jemxadr> Okay, so that produces an AI on all of us to ponder the background pr
+oblem and send all questions (even `stupi' ones) to NJW before the meeting, sa
+y, by Wednesday 20th Jan, and that way the two collators will have time to dis
+cuss the issues
#jemxadr> raised between themselves. END
<huovelin:#jemxadr> Looks fine to me END
<njw:#jemxadr> Agreed END
<allan:#jemxadr> I need a quick reply from as many as possible about a meeting 
+the
<allan:#jemxadr> 1/2 of Feb?
<allan:#jemxadr> END
<Peter:#jemxadr> Fine with me. END
#jemxadr> Yes to Feb 1/2 for me. END
<njw:#jemxadr> 1-2 of Feb is OK with me (actually better than 25 Jan) END
<huovelin:#jemxadr> I think 1/2 Fwb will also suit us well END
<huovelin:#jemxadr> Fwb = Feb END
<allan:#jemxadr> Ok, so much for agenda item 2? END
#jemxadr> I'll push the AI date back to Wednesday 27th Jan for the input of `st
+upid' questions about the background problem. END
<huovelin:#jemxadr> Even better END
<njw:#jemxadr> OK END
#jemxadr> You're all invited to my house for a buffet supper on the evening of 
+February 1st in that case. You were all similarly invited for the review meeti
+ng but then we did the chat thing instead... END
<Peter:#jemxadr> Sounds great! Thanks! END
<allan:#jemxadr> The real advantage of real meetings :-)  Thanks CA: END
<njw:#jemxadr> Looking forward to it, thanks END
#jemxadr> So the meeting's shaping up well. on to the next subject: what do you
+ all think of the proposal for the allocation of error codes? Peter has sugges
+ted that these offsets
#jemxadr> be enshrined in a header jemx.h, and I agree? Shall we formalise this
+ now and then that's one less niggly detail to worry about later. END
<njw:#jemxadr> No problem. But are we all coding in C? END
<huovelin:#jemxadr> We hope so END
#jemxadr> I'm coding in C. Is there anyone using fortran? END
<njw:#jemxadr> I am END
<njw:#jemxadr> Repeat: I'm coding in C END
<Peter:#jemxadr> I'll be coding in C
<Peter:#jemxadr> END
<allan:#jemxadr> Seems that C is getting popular! END
<Peter:#jemxadr> This reminds me: Stephane asked if we could all agree to code 
+in C. In this case he'd not overly bother with providing F90 wrappers for any 
+DAL3 routines he's going to write. END
#jemxadr> I've always had a special attachment to it since it's the first lette
+r of my name!
<allan:#jemxadr> To Peter: I think we can do that for now, but I would not
#jemxadr> So we're all agreed on C? And the offsets suggested in my email can b
+e put into a header jemx.h? END
<huovelin:#jemxadr> I C END
<allan:#jemxadr> like to make a firm committment, we don't know what the
<allan:#jemxadr> future brings with respect to undermanning of the project
<allan:#jemxadr> and subsequent putting people in, who can *only* do fortran.
<allan:#jemxadr> But I see no need for a F90 wrap or similar until that hapens.
<allan:#jemxadr> END
#jemxadr> Are there still people who can *only* do fortran? Shouldn't they expa
+nd their horizons? END
<allan:#jemxadr> They should, but we may be too late for that at the time END
<Peter:#jemxadr> OK. Let's keep it like that then. As long as you don't tell St
+ephane he won't bother about F90 for JEM-X. END
<allan:#jemxadr> Fine, END
#jemxadr> That seems like a reasonable saving of manpower to me END
<allan:#jemxadr> OK next point: END
#jemxadr> We're coming up to the one-hour mark on this meeting, so to cut the o
+ther items a little short, could I just mention that you should all still be t
+hinking of routines
#jemxadr> that would be suitable for the JEMX part of DAL3, and look at Stephan
+e's DAL3 page to get some ideas from his work on SPI.
#jemxadr> Also, the Data level 1 and level 2 datat stuctures are not approved y
+et, so please take a look at them before they get carved in stone, probably ne
+xt week. END
#jemxadr> I still have some things to clear up with Niels Lund on this last ite
+m, so there's time for other considereations here as well. END
#jemxadr> Also something you may have missed is that we now have adevelopers fo
+rum on the SDASt part of the dsri/integral/doc homepage. All contributions to 
+this page, including link URLs to your own work should be sent to me for inclu
+sion.
<Peter:#jemxadr> Will do as soon as I have something (ahem) END
#jemxadr> The page is pretty sparse at the moment, but it does include a questi
+on and answer section which contains all my `stupid' questions to ISDC and all
+ their patient replies. I hope this will be useful to the rest of you. Thats a
+ll my annouonce
#jemxadr> ments, what would other people like to discuss? END
<Peter:#jemxadr> Are you going to write the first version of jemx.h and put it 
+on the WWW? END
<njw:#jemxadr> I'm about to have one or two programs ready that illustrate the 
+use of DAL; I'll hand the URL to CAO soon END
#jemxadr> Yes, I'll make a tentative start on jemx.h, and you can all send the 
+things you think should be included. END
<Peter:#jemxadr> One more point: I'm currently trying to gather some guidelines
+ on "how to write nice ISSW" from our various system people and Stephane (for 
+DAL usage).
#jemxadr> How will this be different from CTS? END
<Peter:#jemxadr> We'll discuss this at ISDC early next week and I'll try to inf
+orm you about the results of our discussions. If you don't hear from me by Tue
+sday evening, just bug me. END
<Peter:#jemxadr> Using CTS and the sample***.c files given is a good way to sta
+rt. 
<Peter:#jemxadr> But one can argue, e.g., about the 'best' way to use groups an
+d the corresponding DAL functionality.
<Peter:#jemxadr> An example: if I have 4 different data structures used by an e
+xecutable and they are all part of the ScW group than it would suffice in prin
+ciple to give the ScW ID to the esceutable as parameter, DAL could find all th
+e input structures from this.
#jemxadr> That sounds nice and handy. END
<Peter:#jemxadr> But Bruce O'Neel argues that we might have special situations 
+where soem structures aren't where they should be and thus we should have one 
+parameter per input and out datastructure - with the option to usually keep mo
+st of them 'hidden'. END
<Peter:#jemxadr> There are some more things. I'll try to have a discussion on M
+onday or Tuesday and will let you know after that. OK? END.
#jemxadr> If a structure isn't where it should be, isn't this a fault in the ar
+chiving system, or whatever puts files where they should be. Or do you mean th
+at one would not be using the default files for some personal reason? END
<huovelin:#jemxadr> OK END
<Peter:#jemxadr> To CAO, the latter. Also we might want to be able to 'fix' a p
+ipeline for QLA quickly even though some data storage is going wrong. END
#jemxadr> We'll wait for your fuller comments Peter, after your discussions. I'
+ll put an action item on you to keep us informed before next friday, okay? END
<Peter:#jemxadr> OK END
#jemxadr> Any other business? END
<Peter:#jemxadr> Not from my side END.
<allan:#jemxadr> No END
<huovelin:#jemxadr> No other business END
#jemxadr> What time for next week's chat? END
<huovelin:#jemxadr> Same time ? END
<njw:#jemxadr> NO other questions END
<Peter:#jemxadr> 9:30 is fine with me END
<njw:#jemxadr> 9:30 is fine END
#jemxadr> Are there people who prefer 9.00? Maybe those a little further east..
+.END
<huovelin:#jemxadr> It's nicer to wake up well beforew the meeting, so lets kee
+p the 9:30 CET END
<njw:#jemxadr> Bye for now  - have a nice week END
*** Signoff: njw ()
<huovelin:#jemxadr> Same to you , bye bye END
#jemxadr> Okay, so next meeting 9.30 Friday. See you all then. Bye! END
*** Signoff: huovelin ()
*** Error: Closing Link: oxborrow[~oxborrow@tenma.dsri.dk] () 
*** IRC session log ended at Fri Jan  8 10:51:18 1999.
