Minutes of Face-to-face meeting 2005-06-23/24

The second face-to-face meeting of the Network was held at ESO, Garching, June 23-24, 2005. The minutes of the meeting are give below.

Participants: VadimBurwitz, NicolaCaon, BiancaGarilli (part), PrebenGrosbol, PeterLinde, DanielPonz, PhilippePrugniel, ChristianSurace, ThijsVanDerHulst, RichardHook, MarkoUllgren, PalleMoller (part), JosephSchwarz (part), DougTody, JohnDavies (part)

Thursday, June 23, 2005 (Chair: PrebenGrosbol)

10:00-10:15 Welcome

• After a short welcome, the agenda was discussed. The 'Update of related project' was shifted to before Lunch. It was agreed that the discussion on the high-level requirement should have priority to ensure that the main discussion of them could be finished during the meeting. Most of the discussions on architecture were shifted to Friday.

10:15-10:30 OPTICON and its future plans (John Davies)

• JD, OPTICON project scientist, gave an overview of the OPTICON activities and its plans (see OPTICON presentation).
  • PG asked on the status of FP7 but it is still too early for details. It is expected that most of the areas supported in FP6 also will exist in FP7 in one or the other form.

10:30-10:45 Updates on related projects

• RH gave an update on SAMPO:
  • The project started in January 2005 and has produced a layer interfacing MIDAS to the Python scripting language. PyMidas makes MIDAS commands available from Python and will be demonstrated at ADASS 2005 as a proto-type.
  • A final plan for the next developments will be made during the fall. In this context, recommendations made in this Network will be included in the considerations.
• JS summarized ALMA related developments:
  • ALMA will base its offline data reduction system on AIPS++ modules. To integrate it well with ALMA, an interface to AIPS++ tasks will be established using CORBA and ACS. For standalone usage it is expected that a 'small' version of ACS will be availabel. ACS provides containers for Java, Python and C++ which could be used.
  • PG reported that ALMA was willing to provide a detailed account of the ALMA/AIPS++ work at the next face-to-face meeting.
• DT reported on relevant projects at NRAO:
  • A framework for offline data processing is considered to support both ALMA and EVLA. The time scale is likely to be a few years. Two tracks are followed namely: a) project driven, with limited scope and hard deadlines and b) a longer term, more research-oriented effort addressing the general problem. Currently two prototypes are in the works i.e. a) ALMA offline, based on AIPS++ modules (now known as CASA) b) integration of AIPS tasks with Python, coordinated also with Radionet.
• DT overviewed recent VO efforts:
  • There is significant work on the definition of the Data Access Layer for images, spectra, and time series data. Proto-type code is available but reference implementations are first expected in a year. The lines between the OPTICON effort and VO work must be clearly defined especially with respect to the AstroGrid where some similarities exist although the scope is different.

10:45-11:00 Revisit of Vision statement, Scope and Use Cases

• At last face-to-face meeting, a few changes to the VisionStatement were proposed. PG summarized the modifications made which were accepted without objections.

11:00-12:30 High-level requirements I (PrebenGrosbol)

• During the previous weeks a subgroup (BG, PL, PG and DP) had detailed the high-level requirements (see DetailedHighLevelRequirements). They were presented for general comments and discussion. Before the individual requirements were considered, three general issues were discussed:
  • Dissemination: Once the requirements are stable they should be brought to the attention of groups working on data processing. It would be useful to create a document which people could refer to. It would help such groups in defining a suitable subset for their project.
  • Verification: It is important that all requirements can be verified.
  • Priorities: There are very many requirement so it is unlikely that all will be implemented. Thus, priorities should be attached to each one.

12:30-13:30 Lunch

13:30-14:30 High-level requirements I (PrebenGrosbol)

• The discussion of the requirements on Installation and running was continued.

14:30-15:30 High-level requirements II (PeterLinde)

• The second group of requirements on Scripting and execution was discussed.

15:30-16:00 Coffee break

16:00-17:30 High-level requirements III (DanielPonz)

• The requirements on Data structures and handling were then considered. Since there had to be time for the contribution of MU, the discussions were stopped around 17:30 - to be continued the following day.

17:30-18:00 Options for container implementations (MarkoUllgren)

• MU gave an account on possible container implementations. The terminology was slight different from that used in the Architectural Concept document. Due to lack of time, the general discussion was postponed to the following day.

Friday, June 24, 2005 (Chair: PrebenGrosbol)

09:00-10:15 Architectural concepts (DougTody)

• DT compared our architectural concept with related VO and Grid work, and reiterated some aspects of the design.

10:15-10:30 Environment levels (PrebenGrosbol)

• The proposal to define explicit levels for implementations of the environment was briefly overview. There was general agreement that it was useful as not all implementation could be expected to support all features.
  • The VO related features at Level 2 should be moved up to Level 1.

10:30-11:00 Coffee break

11:00-12:30 Parameter model (DougTody)

• DT had just before the meeding provided a TWiki page on the revised parameter model (see ParameterModel). He summarized the proposal which then was discussed.

12:30-13:30 Lunch

13:30-14:30 High-level Requirements III (DanielPonz)

• The last part of the Data structure and handling requirements was discussed. The detailed consideration of requirements on Visulization was postponed as time was short. It was also felt that they should be taken together with requirement related to standard utilities and general applications which still have to be formulated.

14:30-15:00 Planning for next 6-9 months

• The following topic and issues should be considered:
  • The next face-to-face meeting should be held just before ADASS 2005 Sep.30/Oct.1. Although there is only 3 months,it will be an advantage since many in any case would participate in ADASS.
  • After the requirements have been updated, they have to be turned into a proper document. This should include a preample describing the background and intend. Also a short description and/or justification should be attached to the individual requirements.
  • An XML definition of the data structures corresponding to the ParameterModel should be made.
  • Work on the specification of subsystems and interface between them must be initiated.

15:00-15:15 Conclusions

• The main objective of the meeting had been achieved as a detailed set of high-level requirements had been discussed and agreed on. This constitutes the first major milestone for our Network.
• The parameter model was revised and now seems to satisfy all basic requirements so that proto-type implementation could be considered.

Please correct my errors and add what I missed. My notes were not all that explicit so I had some problems interpreting them.

Assigned to	Due date	Description	State	Notify
PrebenGrosbol	Sat, 11 Jun 2005	Request information on how will participate in our third f2f meeting	closed		edit
DougTody	Fri, 29 Jul 2005	Provide viewgraphs of presentation at the face-to-face meeting in PDF format	closed		edit
MarkoUllgren	Fri, 29 Jul 2005	Provide viewgraphs of presentation at the face-to-face meeting in PDF format	closed		edit
PrebenGrosbol	Fri, 29 Jul 2005	Update high-level requirement page	closed		edit
PrebenGrosbol	Fri, 2 Sep 2005	Create document with detailed requirements	closed		edit
PrebenGrosbol	Fri, 29 Jul 2005	Submit abstract for poster at ADASS	closed		edit
PrebenGrosbol	Fri, 19 Aug 2005	Request BoF? at ADASS	closed		edit

-- PrebenGrosbol - 22 Jul 2005