Phone Meeting 2005-03-22 15:30 UT

Participants: BiancaGarilli, PrebenGrosbol, PeterLinde, BillCotton, RichardHook, MarkoUllgren, JosephSchwarz, KariVasko

Agenda and minutes

1. General information
   • Changes in the membership list
     • Several changes on the NetworkParticipants page were made. They include adding ChristianSurace and JosephSchwarz, replacing JanneIgnatius with MarkoUllgren and swapping BiancaGarilli and FabioPasian.
   • New phone conference system
     • This time we used a new phone conference system for which it is possible to schedule meeting with very short notice and with little overhead. A side-effect is that the conference first starts when the lead (normally PrebenGrosbol) has phoned in so if you do it earlier you will be put on hold (with music I'm told).
     • PeterLinde asked if it was planned to start the next set of discussions on the requirements possibly using this system. PrebenGrosbol confirmed that it would be good to start these discussions in April. He will over the e-mail list ask who is interested in joining in the requirement discussions and start a set of phone meetings late April (see action).
2. RadioNet software plans
   • Comments on the RadioNet software
     • BillCotton gave a few additional remarks but will send some more details on the fase e-mail list (already done).
3. SAMPO project
   • Short update on SAMPO
     • RichardHook gave a short account. The SAMPO project was started Jan. 2005 with a first small project to make MIDAS accessible from Python. Also time was used to look into the ESO CPL system and VO. In the last weeks several meeting of the advisory committees were held. The Finish committee suggested that the OPTICON architecture should be considered but that one should try to employ an existing, open source container technology. The ESO advisory group will first meet on 2005-03-23 after which the SAMPO project will consider the next steps depending on the different recommendations.
     • MarkoUllgren emphasized that implementing a container from scratch is a significant task and to adopt an existing, open source container technology would be more efficient. The OPTICON requirement on multi-language support was brought up as many existing containers are language specific.
     • JosephSchwarz mentions that the ALMA project had already developed Execution Framework (ACS/CORBA technology) and multi-language containers (for Python, Java and C++).
     • BillCotton and RichardHook expressed concern that a full ACS/CORBA implementation would be too heavy for a standard desktop environment and could turn potential users away. JosephSchwarz replied that light versions had been developed in Java for the observation preparation and soon for C++ for use with the ALMA offline AIPS++ system.
4. Architecture
   • Discussion on minimum container-component interface
     • PrebenGrosbol gave a short summary of his fase e-mail on the minimum container-component interface. There were some confusion on the use of the term 'environment' in the e-mail. It was meant as our Future Astronomical Software Environment and not the operating system environment.
     • BillCotton was concerned with the frequent reference to threading since many processing tasks would be heavy and therefore better suited for multi-tasking with standard processes than for threading. PrebenGrosbol agreed but argued that using a thread type 'run()' method would offer both option as it would be trivial to make a C-type 'main()' task which could run as an independent process and just execute the 'run()' method of the component.
     • JosephSchwarz raised the data model issue on transparent access to different data. It is clear that very different types of data cannot be exchanged without a detailed knowledge e.g. ALMA raw visibility data are simply not compatible with optical images or spectra. On the other hand, it would be nice that image data from the ALMA pipeline would be mapped to e.g. VO compatible formats so that general applications could access both such data and normal images.
     • BillCotton remarked that although radio interferometer images would share many or most of the attributes of optical or infrared images their noise model is significantly different.
5. Actions
   • Status of open actions (see ActionList)
   • Summary of new actions (see below)
6. Time for next phone meeting
   • Suggestion: 2005-04-19 14:30 UT (16:30 CET)
7. Any other business
   • • At next phone meeting we should fix the dates for the face-to-face meeting

Please add what I forgot and correct my errors.

-- PrebenGrosbol - 24 Mar 2005
to top