These minutes are still under construction; the finals are bound to be different.
Day | Time | |||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Mon | 09:00 | Topic | Jürgen Knobloch: Introduction, Agenda | |||||||||||||||||||
References | N/A | |||||||||||||||||||||
Summary |
Workshops in the afternoon:
|
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
09:20 | Topic | Sverre Jarp: RD47 (PCSF a.k.a HEPPC) | ||||||||||||||||||||
References | Transparencies: Web | |||||||||||||||||||||
Summary |
One can LSF jobs from Unix to PCSF soon. This would be a workaround for the impossibility to submit jobs to this farm from outside CERN. ZEUS failed because realy only 1 person worked on it and there was too much time pressure. |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
09:50 | Topic | Jürgen Knobloch: ACOS report | ||||||||||||||||||||
References | Transparencies: Web | |||||||||||||||||||||
Summary |
New members: Helge Meinhard (domain manager), Ruediger Voss (muons), Maya (new librarian). Shows the agenda of the latest ACOS meeting. |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
10:05 | Topic | Jürgen Knobloch: FOCUS report | ||||||||||||||||||||
References | ||||||||||||||||||||||
Summary | Discusses what was discussed in the latest meeting |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
10:20 | Topic | Laura Perini: LCB report | ||||||||||||||||||||
Background | LCB Web pages | |||||||||||||||||||||
References | LCB minutes: Text | |||||||||||||||||||||
Summary |
Overview of the LCB.
RD45 (referee Bob Jones).
ROOT
LHC++
GEANT4
Videoconferencing
Members of the LCB from ATLAS are no longer ad personam but represent the experiment. They are supposed to present and discuss agenda items for the LCB in the appropriate ATLAS meetings beforehand. |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
10:50 | Topic | Chris Onions: Software Release Tools | ||||||||||||||||||||
Background | See CVS working group page | |||||||||||||||||||||
Decisions to be made | What is the deadline for having all existing packages integrated into the CVS repository? | |||||||||||||||||||||
References | N/A (ask slides) | |||||||||||||||||||||
Summary |
Rosemary asks for a clearer description of responsibilities. Who may check in new things? Domain responsibles. BaBar has some 160 "packages", we'll have something similar. Daq people are also using it and wrote an introductory user guide. A full manual is in the make by Lassi. |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
??:?? | Topic | Helge Meinhard: GNATS | ||||||||||||||||||||
References | Transparencies: Web | |||||||||||||||||||||
Summary |
|
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | Start thinking now about which categories we have and who are responsible. | |||||||||||||||||||||
??:?? | Topic | Helge Meinhard: DIG report | ||||||||||||||||||||
Background | See the minutes of the meeting on 03/07/1997, and the agenda of the meeting on 28/08/1998. | |||||||||||||||||||||
References | Transparencies: Web | |||||||||||||||||||||
Summary |
DIG=Domain Interface Group. Related to OO software. Other relevant group is the support team.
Domains have been defined and have started design, but this first decomposition will almost certainly change. Domain teams nominate their representative in the DIG together with the other responsibilities. |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
14:00 | Topic | Extra DIG Meeting | ||||||||||||||||||||
Present | Helge, Patrick, KB, Maya, Steve, Jim, Dave Mellon, Christov, RD, Julius, Roger, Toby | |||||||||||||||||||||
Summary |
Magnetic Field
Domain Decomposition
Integration of Graphics into ARVE
Utilities domain
Redefine MuonSystemPropose to have a meeting on Wednesday 17:00 to discuss the MuonSystem review. ReconstructionDiscussion on new OIG's with Patrick Hendriks, Roger Clifft, Kors Bos and anybody else from the Reconstruction domain on Tuesday 17:00 hrs. Control
InnerDetectorSystemA meeting will be organized by Roger with Pop tomorrow to see where we are and see how we can make progress and to try to coordinate the work going on by various people like Peter Luthaus, Jim Loken, Armin Reichold, Alan Poppleton. Java
SLUG Class
Note added later: this meeting took place and Andrea dell'Acqua agreed to implement this class. |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions |
|
|||||||||||||||||||||
16:00 | Topic | ASP meeting | ||||||||||||||||||||
References | ||||||||||||||||||||||
Summary |
|
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions |
|
|||||||||||||||||||||
Tue | 09:00 | Topic | Steve O'Neale: SLUG/DICE/ATLSIM consolidation | |||||||||||||||||||
Background | Steve O'Neale's consolidation page. | |||||||||||||||||||||
References | Transparencies: Web
Consolidation Web pages, ATGEANT3216 |
|||||||||||||||||||||
Summary |
The objective of the consolidation effort is to produce a single package incorporating the batch and interactive features of SLUG, DICE, ATLSIM, together with GEANT(Atlas), GENZ, GENSLUG, GGENZ, ATRECON, ATRIG and the CERNLib libraries and code of 97a plus corrections... The adoption of tools such as SRT, CVS, cpp etc especially since CMZ is no longer supported is essential in order to facilitate the work of the programmers and the users. The resulting package, with a new name and structure to be decided upon, will be in the SLUG framework. ATGEANT3216 has been built based on the latest GEANT3 release (version 8) and includes several ATLAS changes (see ATGEANT3216 ). Experience with cpp in compilation and library installation has been discussed with the SRT/CVS working group. The consolidation of the ATLSIM package involves as a first step translation into FORTRAN and cleanup of duplicated code (see also ADA's transparencies). The treatment of pileup where ATLSIM is faster and of random number seeds and initialisation where general improvement is needed were presented. Extensive testing by the user community is required. An overall revamping of the documentation and subsequent documentation support are urgently needed. | |||||||||||||||||||||
Discussion |
A. Amorim said that a `datacard database' scheme (probably initially on WWW) is important for testing. Such a scheme exists for ATGEN datacards but further user seems to be hampered by the user `psychology'... An automated book-keeping system for all user jobs was also advocated. J. Knobloch emphasised that an update of the list of supported platforms in necessary. CERN-based repositories for non-supported platforms may still be available. |
|||||||||||||||||||||
Decisions | ||||||||||||||||||||||
Actions |
|
|||||||||||||||||||||
09:00 | Topic | Andrea dell'Acqua: SLUG/DICE/ATLSIM consolidation | ||||||||||||||||||||
References | Transparencies: Web | |||||||||||||||||||||
Summary |
The aim of the consolidation project is to bring the current software into a coherent and maintainable state so as to allow the completion of the physics TDR and ensure a smooth transition to the OO software. One of the main issues is the merging of the mainly interactive ATLSIM program with the SLUG/DICE batch framework. A beta version of the resulting interactive/batch executable could be envisaged for end of September. | |||||||||||||||||||||
Discussion |
The definition and purpose of the consolidation were briefly discussed following a question raised by K. Sliwa. A. Amorim proposed the creation of a user `job database' for mapping the various uses of the software. The purpose and feasibility of this were debated. It was noted that an effort to collect the user requirements/expectations has already been carried out by J. Pater et al in the Inner Detector community. This has served as input to the consolidation effort. The dependence of the quality of the physics results on the software quality was emphasised. A measure suggested was to require the submission of the software used to obtain physics results together with the note in which these are published. At this point, it was noted that the compatibility of the two current software frameworks in terms of physics has so far largely been maintained. |
|||||||||||||||||||||
Decisions | ||||||||||||||||||||||
Actions |
|
|||||||||||||||||||||
09:45 | Topic | Gilbert Poulard: Status of 97_6 release | ||||||||||||||||||||
References | Transparencies: PostScript | |||||||||||||||||||||
Summary |
The 97_6 version of the ATLAS software is the last CMZ release before the migration to CVS. It is essentially complete in terms of geometry (the muon code has been added) and reconstruction for the calorimeters and the Inner Detector. The trigger code (ATRIG) and the muon recontruction code remain to be added. An ATRIG/ATRECON executable has been requested by the physics community. | |||||||||||||||||||||
Discussion |
A. Amorim suggested an update to the latest version of PYTHIA. C. Onions pointed out that the inclusion of new code should be in the CVS rather than in the CMZ framework to avoid delays. G. Poulard argued that a CMZ version should not be too difficult provided it does not delay the migration. |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
10:30 | Topic | David Candlin: Arve ramp-up evaluation results | ||||||||||||||||||||
References | Transparencies with comments: PostScript, DVI, TeX | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
10:50 | Topic | Toby Burnett: Arve ramp-up status | ||||||||||||||||||||
Background | See Toby Burnett's status page. The ramp-up is based on the results of the evaluation (see reports by Lassi Tuura and David J. Candlin) and the discussion in the Domain Interface Group (see minutes of meeting on 03 July 1997). | |||||||||||||||||||||
References | Transparencies: Web | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
11:30 | Topic | Lassi A. Tuura: Arve ramp-up timescale | ||||||||||||||||||||
References | Transparencies: Image | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
13:30 | Topic | A short meeting to identify improvements in ATLAS computing welcome and documentation, for newcomers. | ||||||||||||||||||||
References | N/A (ask slides) | |||||||||||||||||||||
Summary |
|
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
14:00 | Topic | Database meeting | ||||||||||||||||||||
Background | The G3/Zebra to G4/Objy converter should be an issue. | |||||||||||||||||||||
References | Agenda, Minutes | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
Wed | 09:00 | Topic | Jamie Shiers: LHC++ | |||||||||||||||||||
Background |
Jamie Shiers has been asked to report on the status, plans, and timescales for
We'd also like him to discuss
|
References | Transparencies: Web | |||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
11:00 | Topic | Bob Jones: ATLAS DAQ Software Process | ||||||||||||||||||||
Background | The main idea is to receive input for the Editorial Committee for the Atlas Software Process ASP. | |||||||||||||||||||||
References | Transparencies: PostScript | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
11:30 | Topic | Katsya Amako: GEANT4 Software Process | ||||||||||||||||||||
Background | The main idea is to receive input for the Editorial Committee for the Atlas Software Process ASP. | |||||||||||||||||||||
References | Transparencies: Web | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
14:00 | Topic | Graphics meeting | ||||||||||||||||||||
References | Agenda, Minutes | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
Thu | 09:00 | Topic | Hans-Peter Wellisch: GEANT4 Framework | |||||||||||||||||||
References | Transparencies: PostScript | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
09:30 | Topic | Andrea dell'Acqua: GEANT4 ATLAS Implementation | ||||||||||||||||||||
Background | A definition of hits and digis objects is required such that work on the G3/Zebra converter to G4/Objectivity can start. Pile-up to be covered. |
|||||||||||||||||||||
References | N/A (ask slides) | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
10:30 | Topic | Maya Stavrianakou: Productions and LSF | ||||||||||||||||||||
References | Productions transparencies: Web
LSF transparencies: Web |
|||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
11:30 | Topic | Sverre Jarp: Cocotime Request | ||||||||||||||||||||
References | Transparencies: Web | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
11:45 | Topic | Steve Fisher: StP, CodeCheck | ||||||||||||||||||||
References | N/A (ask slides) | |||||||||||||||||||||
Summary |
Steve told us about StP version 2.4 (beta release) which he has been testing on Windows NT. It has UML, but he has not yet tried it. He suggests that we should not change to UML too soon. The PC version of StP is very similar to the Unix version. You are provided with MKS (Unix style tools), a frame viewer, Sybase, an X server and FlexLm. It is designed to inter-work with Unix. For example Sybase could be on a Unix machine. A significant change to this version is that it no longer uses SNiFF+ as a parser for the class capture, but uses their own fuzzy parser. The comment capturing does not seem too good, so we make have to continue using the existing home made script for this. A few small changes have been made to the customisation of StP. A report generator to extract headers from requirements tables to generate a decomposition document has been written. It is now possible to exclude diagrams from a report and to control the displayed version number of a report. He then told us about CodeCheck, which is a parser for C++ which makes it very easy to write programs to check coding rules or to generate metrics. Many example programs are provided with the tool. It works in command line mode on Unix or PC, but can also be integrated with Visual C++ as a tool allowing one to jump quickly to code rule infringements. A program (of about 800 lines) allows 37 of the 79 Atlas Coding Rules to be checked. The others are either subjective or too hard to check with the current version of CodeCheck. A survey of recent work on OO metrics revealed that not many metrics had been validated, and that many validations had been carried out on student projects. Steve said that he was especially interested in metrics which can be extracted from the design and proposes to add these to StP design reports. He had looked at three metrics for implementation, Decisions per executable line, Operators per executable line and number of executable lines of code. He showed a distribution of executable lines of code for 147 functions, which indicated very small functions (about 4 lines). Design metrics considered were the depth of the class in the inheritance tree, the coupling between classes (number of distinct classes referenced) and the counts of public, protected and private methods. It was noted that the count of private methods could not be made at design time and the count of coupling between classes would depend upon whether the measurement was made at design time or from the code. Coupling for 31 classes he had looked at showed a peak at around 2. Finally he observed that there were still problems with the parser being unable to cope with some STL, but that the company fixed bugs very quickly and that he hoped to soon be able to recommend purchase of the product. |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
12:00 | Topic | RD Schaffer: Database Working Group Report | ||||||||||||||||||||
References | Transparencies: PostScript | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
12:00 | Topic | World Wide Computing meeting | ||||||||||||||||||||
References | Agenda, Minutes | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
Fri | 09:00 | Topic | Krzysztof Sliwa: Word Wide Computing Workgroup Report | |||||||||||||||||||
References | Agenda, Minutes | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
09:20 | Topic | Julius Hrivnac: Graphics Workgroup Report | ||||||||||||||||||||
References | N/A (ask slides) | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
09:40 | Topic | Kors Bos: Summary of Decisions | ||||||||||||||||||||
References | N/A (ask slides) | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
10:10 | Topic | Jürgen Knobloch: Planning for Next Cycles | ||||||||||||||||||||
References | N/A (ask slides) | |||||||||||||||||||||
Summary | ||||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
11:00 | Topic | Helge Meinhard: Dates of 1998 Workshops | ||||||||||||||||||||
Background | Must take into account
One workshop should be held outside CERN (US?) |
|||||||||||||||||||||
References | N/A | |||||||||||||||||||||
Summary | Tentative 1998 calendar: 23/02 - 27/02 Software workshop, CERN 02/03 - 06/03 Atlas week, CERN 28/03 - 04/04 Physics workshop, Grenoble 04/05 - 08/05 Software workshop, CERN 08/06 - 12/06 Atlas week, CERN 24/08 - 28/08 Software workshop, US (alternatively: 07/09 - 11/09) 31/08 - 04/09 CHEP conference, Chicago 07/09 - 11/09 Software workshop, US (alternatively: 24/08 - 28/08) 14/09 - 18/09 Atlas week, CERN 16/11 - 20/11 Atlas week, CERN 07/12 - 11/12 Software workshop, CERN |
|||||||||||||||||||||
Decisions | N/A | |||||||||||||||||||||
Actions | N/A | |||||||||||||||||||||
??:?? | Topic | Working Group Meeting on Testing | ||||||||||||||||||||
References | Transparencies: PostScript | |||||||||||||||||||||
Summary | The slides briefly list:
|
|||||||||||||||||||||
Discussion | It was not possible to come up with definitive plans without
further work, but there was a lot of usful input about things we might
look at. These are some of the topics.
There was also some discussion of the extent to which we could follow industry practice and keep testing independent of implementing. It was felt that on the whole we had too few people to be able to do this, and from a practical point of view, most of the tests had to be devised by the same people as wrote the code. However, we may be able to persuade small groups of people to swop tests between themselves. There is also a role for independent testers (who may be volunteer users) to do spot checks or to test especially important parts of the code. There was little enthusiasm for the "bazaar" method of developing code (more or less what we had at the moment!) because physicists in general do not form a pool of enthusiastic code hackers and want their code to be of good quality as soon as it is released. |
|||||||||||||||||||||
Decisions |
|
|||||||||||||||||||||
Actions |
|