From: Configuration Control Board

CCB Meeting 04 June 2008

Attendees

Robyn Allsman
Tim Axelrod
Roc Cutri
Ray Plante
Robert Lupton

Kian-Tat Lim, a proposal originator, was unable to attend this meeting.

Proposals Discussed

Meeting

Typedef Proposal

Discussion orchestrated by Robyn.

Non-controversial proposal was accepted by CCB board. Robyn changed the C++Standard document.

Bottom Left Pixel Proposal

Discussion orchestrated by Robert.

Roc and Tim were concerned that the proposal didn't provide adequate detail to map the software (0,0) reference point onto individual amplifier rafts. They want a solution which is fixed regardless of the potentially changing raft orientations over the life of the project.

Robert agreed to write and circulate a DM Policy statement on such pixal address issues.

This proposal was deferred until the Policy is written.

Java For Event Monitor

Discussion orchestrated by Ray.

Ray split the discussion into two parts:

  • what criteria are important when considering Java as the implementation language for a DM software;
  • does the Event Monitor satisfy the criteria for selecting a alternate implementation language.

At the moment, the only criteria used when deciding whether to use Java, instead of C++ or python, is if the java will be entirely separate from the core of the DM system. This generally has be taken to mean that Java could be used for UI situations. Ray agreed to writeup the concensus opinion on use of alternate languages in LSST software stick.

Tim suggested that if there is no compelling reason to introduce another implementation language then itshould not be done. Although the project might be tempted by current the availability of a java wizard programmer, gratuitous java implementations just promote longer-term maintenance issues. He suggests that a compelling reason to use Java might be if a respected library or interface exists only in Java and not C++ or python.

Since java doesn't provide an events framework which is better than existing python or C++ libraries, it was not a candidate for java implementation.

This proposal was rejected since it doesn't meet the newly stated criteria for new language use.

Package Proposal

Discussion orchestrated by Robert.

Tim felt the need for a higher level discussion before delving into the nuts and bolts of the proposal. As a non-power developer, he wants a simple interface with the defaults set for the usual case. If a wrapper is needed to provide that simple interface, it should be provided.

The need for better eups documentation and additional documentation on the use of eups by DM was mentioned. Robyn suggested providing boiler-plate scenarios for the most common usage.

Robert will revise the revised Package Proposal, then send to KT, next Tim and Nicole will review it as 'typical' developers. The new document will be reviewed at the next CCB meeting on June 25. The document should be ready for CCB review in advance of June 16 when Ray departs for vacation.

This proposal was deferred to another CCB meeting.

Action Items

  • Robert Lupton will prepare a DM Policy statement on the pixel (0,0) reference origin. The Policy statement will be circulated for review and comment.
  • Ray Plante will prepare a DM Policy statement on the essential conditions for Java to be acceptable as a DM implementation language.
  • Robert Lupton will revise Kian-Tat Lim's Revised Package Proposal. Tim and Nicole will review it on behalf of 'everyman'.