ALMA Software Group Minutes, 2000-Jul-12, 14:30UT (08:30 MDT, 16:30 CET) ESO: Chiozzi, Filippi, Gustafsson, Heald, Karban, Raffi IRAM: Lucas NRAO/CV: Pisano NRAO/SOC: Glendenning, Harris, Stauffer New Discussion Items ==================== 1. Project News (Glendenning, Raffi) Glendenning reported that senior project management met in Charlottesville to discuss Phase 2 task division at a very high level - task separations will be decided later. For software the decision was made to split the effort 50/50 (defined by Phase 2 cost estimates) between Europe and the US. Raffi added that on the European side the institutes are determining their degree of interest in various activities. For the correlator the baseline decision is to use the NRAO correlator design. At about the end of this year a report on the possibilities for a "future" correlator - possibly a European/Japanese collaboration - will be issued. If the additional scientific capabilities offered by that correlator warrant, it will be built with the extra costs coming out of contingency. Probably the 1st quarter of the NRAO correlator would be built for interim operations. 2. Area Reports a) SSR (Lucas) Have received responses from ASAC about questions posed to them by the SSR. The answers correspond to the consensus of the group. More discussions of blanking and related issues has been taking place - there is no consensus yet. Some progress has been made in preparation of the detailed requirements list - a draft should be issued in about a month. Use case work is proceeding. It is clear that they will have to go to the SSR as new points are (not surprisingly) being discovered. Not yet clear how much can be translated into use cases - the process is more complicated than initially thought. b) Analysis & Design (Harris) Providing context and effort for use cases. After the Grenoble meeting effort with Grosbol and Mudders working towards a high-level analysis and design document. Wants to then move into architecture definitions, and then push into more detailed and simplified architecture for the test interferometer. c) Software Engineering (Filippi) Design methodology and tools document to be issued in about 10 days. This version will include material orriginally from the ESO and NRAO documents, as well as comments and discussion that have happened since they were written. Filippi indicated that next a kit would be produced for use by anyone who wants to use the procedures to make the standard more effective. Kit includes things like compilers, TcL, etc. While we are principally using Rose, various tools etc are needed to extract and insert text, etc. Glendenning asked whether we shouldn't just maintain a list of required software versions? Filippi responded that in his experience a kit was very useful. After some discussion we decided to start without such a kit. Within the year an additional person at ESO will start work in this area. Bridger is working on coding standards, but Filippi was not sure of the exact status. Harris asked about defining document types and formats? Raffi responded that we deciding on the types is not a SE responsibility - we need to fix this in our own minds. Standardizing on format might make sense. d) ACS (Chiozzi) Raffi introduced the subject by noting that the quite a few internal comments on the ACS document were received. He stated that we have decided to go to review on this document, marking as TBD areas in which we do not have good experience. Chiozzi carried on mentioning that he had received quite a few comments, some substantive, and some quite late. He intends to circulate internally replies and and updated document. Some items will be marked TBD and postponed to a later revision of the document. He reported that they are doing some real prototyping. BACI - the basis of ANKA - has been installed. Now running under Linux and controlling a "power supply" with a Java interface. Some modifications to ROProperty have been made. There are still some VxWorks compiler problems. They have investigated how to extend the set of ROProperties. Heald has submitted IDL for mount interface, which may be incorporated for the late-year Kitt Peak tests. Stauffer asked if class libraries are available for ROProperty and Property. Glendenning added that these Properties and Characteristics are so fundamental it would be nice to see their interfaces - even in UML or IDL with no executable code. Chiozzi responded that we have the IDLs, in more or less final form. => Chiozzi to distribute IDL and/or UML for Properties and related classes. e) Control (Glendenning) There was an "interfaces meeting" held in Socorro in June. While mostly about electronics interfaces, some action for us came of it. The specification for devices requiring precise timing was tightened. Brooks was asked to (and has) produce a document listing options for total power interfacing. Heald issued an internal document outlining his decision to perform coordinate conversions at the ABM. Stauffer and Glendenning have been working on a document describing standardized states and nomenclature for documenting device drivers [since renamed "device controllers."] This will be sent out more widely soon. The commanding styles for a couple of particular "tricky" devices has been discussed with their designers. In particular for the fringe generator and nutator. The former resulted in a lengthy discussion about the correct was to move the phase centers while doing OTF mosaicing. Brooks and Glendenning have been improving the planning for the control software group. f) Correlator (Pisano) Focussing on test correlator design and prototyping. More slowly than he had hoped. Working through fine-timing details. Prototype work is addressing fine delay issues. Consultation with Steve Scott has been invaluable for working through these issues. 3. KP 12m Availability (Glendenning) Glendenning briefly reported that negotiations for new institution(s) to run the Kitt Peak 12m telescope are ongoing. We remind the people engaged in the negotiations from time to time about our desire for occasional tests on the 12m. There's no known problem with arranging such access, but until there's a final agreement we can't be sure. 4. Antenna PDRs (Heald, Raffi) Vertex PDR on June 19-20. Overall went very well. About 25 points to be followed up. Heald met with the Vertex software people after the PDR. The most interesting discussions were about timining. The 10s circular buffer requirement will be removed. Vertex has agreed to use our pointing model for autonomous mode. There was a lot of discussion about UPS communication and about the touch screen (light levels at high site concern Emerson). There was an amusing error caused by an incorrect units conversion (factor of 10) from English to Metric units: Vertex was proposing to build a 240m^3 concrete foundation! Raffi reported that EIE has not yet held their control system PDR (The general PDR was held). The Vertex and EIE ICD's are diverging somewhat. For example, the antennas are no longer plug compatible (Vertex has 2 nodes). He is working on an exercise to identify the common messages and bring them out into a common ICD section. He will bring this to the European contractor. 5. Face to face Fall meeting (Glendenning) The main constraints are the antenna CDR's, the EIE one is currently set for the week of Sept. 18th which we had previously thought might be a good week. Other possibilities include synchronizing with ADASS, the SSR meeting (date TBD still), or with the KP 12m tests. => Glendenning and Raffi to decide on joint meeting dates. 6. Other topics (All) Karban - Has principally been involved with discussions and previewing the SE document. Lucas - Is writing a document summarizing the IRAM capabilities for data reduction for the test interferometer. With Glendenning is starting work on defining data formats. Continuing Action Items ======================= x> Brooks to circulate the Tucson real-time programmer description once it moves beyond the required internal NRAO advertizing. Done x> Brooks to send Raffi the current implementation guidelines. Done x> Brooks to send Raffi the cable size information. There is some info in Antenna ICD#4 about this. Done x> Glendenning to arrange getting monthly reports from alma-sw-workers membership in advance of our monthly meetings. Done. He thinks that the monthly reports were quite successful - thank you. x> Filippi to report on SE proposal status. Done. ========= Monthly Reports - June ========= Alan Bridger As Giorgio indicated I now should be able to contribute to Alma at about the 25% level, and Giorgio has suggested I first of all pick up on reviewing the C/C++ standards and defining the document templates and document coding system. C/C++ Standards: I have found the proposals from Jim and Mick and have begun looking through them. I'm assuming that all s/w workers should be involved in reviewing them, but I'd like to see what changes (if any) I think might be needed first, and whether or not I feel these documents should be consolidated. Are there particular procedures or timescales for this review (other than ASAP) that I should know of and are there any other reviewers than the s/w workers? Apologies if this is known already, but being a recently joined member I don't yet know everything! Document Formats and Templates: Giorgio indicated a suggestion received from Brian - if someone can point me to that then I can begin work here. -------------------- Mick Brooks - Took part in Test Interferometer planning meeting in Socorro. - Continued design and layout of next AMBSI hardware revision. - Proposed scheme for interfacing total power data to computing system. - Began detailed task planning for NRAO ALMA computing group. - Revised Antenna ICD 9 to include new information from Vertex. -------------------- Gianluca Chiozzi - ACS Architecture Document - circulated for internal review in the ACS mailing list - comments (from M.Brooks and B.Glendenning) being processed - replies to comments + new release to be circulated next week - ACS 0.0 development: - porting of ANKA to our environment fro ACS 0.0 proceeding fast - next week visit to ESO of KGB team for first official installation - check point last week with A.Perrigouard of current status. TCL/Combat tests will be resumed in August, but Combat supports now officially Orbacus - development of test device procceds. Touching up and documenting. Developing prototype of simulation to demonstrate the concept - Kitt Peak test: - Discussed with G.Raffi and R.Heald scope of Kitt Peak test - R.Heald working on Antenna Mount device. Now working on IDL - R.Heald and A.Perrigouard evaluating VLT mount control SW for suitability for ALMA Phase 1. Result of evaluation will affect Kitt Peak test scope (see report on discussion) - XML for ACS - F.Stauffer working on XML issues. He should send you a report -------------------- G.Filippi: ALMA - Software Engineering - Status report 2000/06/28 As the first priority, we concentrated to work on the Analysis and Design and the planned actions are done or close to completion: - agreement with Rational Rose done, first set of licenses (20) bought - first version of the Guidelines, based on Gianluca's work and restructured by Paola will be ready at the end of week 27 (07/07). The target was to have an agreed and workable system, it includes some of the comments that we got from the joint work. As we agreed at the last genearl meeting, this version will be distributed, together with the licenses, templates and software tools, to the people involved in Use Cases and a first review is foreseen for Fall 2000. Good news also on the resource side: I got confirmed a ESO staff position for SwQA as part of the IT department. The person should be operational within 4-6 month from now, giving therefore good chances to a decent support for Phase2. Phase 1 is currently covered with a best effort approach. With some internal replanning, Gianni made available some time from Paola Sivera to support the UML part. Alan Bridger (UK) also gave his availability so he will get some of the procedure (see below). The status of the remaining items (based on the SE Practice document framework) is as follows: - SE Practice document (framework) GIF end of July/mid August - Document Review written procedure GIF end of July - ICD: some work started also as part of the System Engineering (to be coordinated with G.Than). For the Sw part, the proposed Guidelines includes ICD requirements. - Coding standards: received C and C++ proposal. Alan will manage the review. - Configuration Management: a CVS server has been set by Paola for sharing documents (Gianluca's request). Written instructions are also available. We are currently rewriting the CMM for a more robust client-server protocol. The options still are: - to use CVS for documents and CMM for code - to use CVS for everything - Document Formats and Templates: some suggestion received from Brian. One of the possible area for Alan. - Document Numbering: waiting for Brian/Gianni (bring the issue to Project level and decide on a scheme). - Inspections: for Phase 2 - Document Archive: current system is enough for the moment. - Testing: Not discussed. - Fault Reporting: Not Discussed. -------------------- Brian Glendenning 1. With Fritz I worked on generic description of low level devices (including standard states), and a couple of example devices. 2. With Mick I worked on control software planning (primarily for the test interferometer). 3. Phase 2 planning (task division memo to senior ALMA-US management). 4. Many discussions and meetings. Perhaps the most interesting was the Socorro interface meeting and follow-on discussion (e.g. I participated quite a bit in the discussion leading to the M&C precise timing specification). Perhaps also notable were some detailed ACS comments I sent on the -common list. What I did not do: a. Finish the project book chapter. The pressure has not been intense on this item so instead I worked on detailed design w/ Fritz. b. Finish the "Joint SW Plan" document as I've been waiting to see whether the detailed planning would have any effect (it is having some). -------------------- Birger Gustafsson During this period I was on leave for two weeks, so my activity was low. Together with Gianluca I finalized the ALMA Common Software Architecture. I also took part to discussions with Gareth and Ron. We also had a telephone meeting with Mark Plesko. He told us that they are more or less on track. Matej (one of his students) will come to ESO the week 10th of July to install the ANKA software here. I have also continued working on our DO model, now mainly on a manager which will handle simulation transparently to clients. So the client would not know which DOs are in simulation. This will be part of the configuration of the manager. -------------------- Gareth Harris The STATUS TO DATE of my recent activities and current plans on the ALMA project include: 1 - Analysis of the Science Software Requirements [SSR] [in memo 293 and SSR subcommittee meeting May 15..] a - production of SSR Use Cases [SSR UC] [nearing completion of first draft - end July] b - production of SSR UC State and Activity Diagrams [nearing completion of first draft - end July] [ACTIVITY SUMMARY: Met with SSR subcommittee in Garching May 15-18, created first use cases, created first state and activity with swim lane diagrams, held additional conference calls aiming at final merger of cases in Grenoble July 24-28] 2 - Production of High Level Analysis Document [HLA] [combining SSR UC output, ESO DataFlow methods, and my methods of requirement identification] a - initial goals and outline [with Preben Grosbol, nearing first outline July 12] b - first rough draft [goal - August 3] c - first draft with Preben Grosbol and Dirk Muders [goal - August 31] [ACTIVITY SUMMARY: Met with Preben Grosbol in Garching, gathered ESO sample documents - found method similar to mine, created initial goals and document outline, planned further work and meeting in Garching July 31 - Aug 3] 3 - Architecture and Design 1) Overall System Architecture, [One very preliminary overall System Architecture document. PLANS - produce another version in September.] 2) General Design of the final ALMA Control System, [Discussion and agreements with ESO and NRAO on general design aspects such as decoupled control tree and recording stream. PLANS - a simple version in September.] 3) Specific Design for the Test Interferometer, THIS IS THE MOST IMPORTANT UPCOMING ACTIVITY. [In process - basic needed elements identified, PLANS - Detailed document beginning in September.] 4) Technology Investigations and Choices. [Use of CORBA, C++, Java, Python, Jpython, XML, etc. in conjunction with automated creation and configuration of core control tree elements.] [ACTIVITY SUMMARY: This work has more hours but not yet much documentation. Much of the work included: Extensive investigations with Fritz, Discussion with Ron, Further discussion and coordination with ESO in Garching. TESTED: CORBA and distributed classes performance; object classes giving device distribution; maturity evaluation of java, python, jpython, CORBA. test of simplified control tree communications with ESO, integration of jpython, java, and ORB.] -------------------- Ron Heald I've been here at ESO since June 1. I've been having discussions with G. Chiozzi, G. Raffi, B. Gustafsson, K. Wirenstrand, and R. Karban. Topics have been the ALMA Common Software (ACS), the proposed tests to be done at the 12M this winter, and selected subjects from the Glossary, "Device Requirements and Use Cases", and "Antenna Mount Requirements and Use Cases". These talks are a great help in understanding the division of opinion, experience, and culture between our two groups. I attended the Vertex PDR in Duisburg on June 19,20. Most of this was fairly boring for me (no software). The best use of time was a half day discussion with the Vertex antenna drive team (K. Willmeroth and L. van der Horst) on the day following the PDR. I prepared notes from our talks and am now comparing them with M. Brooks (the keeper of ICD#9, the antenna monitor and control interface). The notes can be seen at http://www.aoc.nrao.edu/~rheald/VertexPDRnotes.htm Alain Perrigouard was at ESO for four days, June 26-29. Alain and I are developing the first version of the Antenna Mount Software (AMS). We were able to come to an agreement on what parts of the system will run on the ACC and the ABM. A draft document discussing the decision can be seen at http://www.aoc.nrao.edu/~rheald/AMSPlacement.htm. J. Benson released a source version of CALC to me and I spend some time getting it to run here at ESO. CALC is a package from NASA Goddard that will be used to compute the delays for the array. It produces the antenna azimuth, elevation positions as a byproduct and we could us them to position the antenna. Alain is keen to run some tests on it and will do so when he returns from holidays in August. We have also come to some conclusions on the 12M test. We have set a deadline (August 1) for a decision on including the initial AMS with the test. The decision will be based on how well we can agree on an interface between the AMS and the upper-level control software being developed here at ESO. The interface is being written in IDL using the constructs from CoCoS. -------------------- Roland Lemke What I did in the last weeks was: - performance tests with mySQL database - getting aquainted with CORBA (MICO for CORBA on Linux) (I heard already from Gianluca that MICO will not be used..) - purchased CAN boards and starting to adapt drivers for Linux and RTLinux -------------------- Robert Lucas Done: - Received response from K. Menten (ASAC) which is very much in line with the consensus in the SSR committee. - More discussions on issues of blanking, flagging. - Use Cases meeting in Grenoble (24/28) Next: - Phone meeting later on July 12th - Prepare data format content for test interferometer - Summarize IRAM software capabilities for the Test Interferometer data reduction (instrument calibration) - Use Cases phone meeting next week (July 11) - Use Cases meeting in Grenoble (24/28) Problems: - Scope of Use Cases vs Requirements. -------------------- Status Report for June - Jim Pisano 1. Continuing to update the ALMA Test Correlator Control Computer Software Design. 2. Continuing prototype work on the above software design, specifically with fine delay adjustments, sub-integration (a.k.a. correlator dump) blanking and queuing, and all 800MHz correlator modes. 3. Clarifying timing issues for the test correlator as the basic hardware timing of the test correlator is incompatible with the prototype correlator and the timing events of the ALMA array. -------------------- Gianni Raffi - interactions about critical points for European antenna PDR - organization of control PDR with EIE for 21st of July. - discussions on ALMA design, ACS Architecture and Control software (with Gareth, Ron and Alain during their visits and with Gianluca) - ESO informal meeting on Control model and Control sw for Kitt Peak test - participation to a few telecons (interesting the one on Use Cases) - internal ESO discussion on ALMA sw partecipation in relation to other ESO projects. --------------------