May. 23, 2008 Software
- Review minutes from February 8, 2008 Software meeting
- Recent changes to JANA (David L.)
- Geometry access in DANA (David L.)
- Supported Platforms Discussion
- Action Items
The meeting will be at 1:00pm in Cebaf Center F326-327.
To connect by telephone: 1.) dial:
800-377-8846 : US 888-276-7715 : Canada 302-709-8424 : International
2.) enter participant code: 39527048# (remember the "#")
We will attempt an ESNet connection for this meeting using the following number:
We have a reasonable expectation that this will work.
Attendees: David L., Matt S., Elton S., Alex S., Elliott W., Simon T., Mark I., Beni Z.
Recent Changes To JANA
David presented some slides on recent updates to the JANA framework. These included adding associated objects to existing data objects, incorporating autoconf in the build procedure, and revamping the entire mechanism for printing single data objects to the screen.
It was mentioned during the presentation that the new "toStrings" methods provide the possibility of doing some type of automated I/O of objects. Elliott noted advancements in the BOOST library regarding object persistence that my be useful. He and David would discuss it offline to see if any advantage could be gained from there.
Geometry Access in DANA
David presented some slides on recent updates to DANA (and JANA) that now allow access to the XML-based geometry info. from within the reconstruction code. The XML itself can be queried via an XPATH-like syntax. Because the "view" of the data represented in the XML is geared toward GEANT, it is not necessarily the optimal view for the reconstruction code. The example of the z-locations of an FDC wire plane was given. This number is a sum of several locations from the nested geometry structure represented in the XML. Therefore a single XPATH could not be specified to find it. To address this, a set of "convenience" methods have been set up in the DGeometry class that encapsulate much of the specific knowledge about what structure of the geometry is in order to extract the appropriate numbers and combine them into the values useful to reconstruction.
Some discussion was have with suggestions on possible alternatives to the "convenience" methods. These where:
- From Beni: hdgeant already produces a ROOT file with geometry info in it, including the relationship between volumes. One might be able to use this along with existing ROOT methods to extract the desired info using a single call (again, think FDC wire plane Z-position)
- From Elliott: Instead of doing the conversions of these numbers every time the program is run, one might be able to convert the geant-biased XML into a reconstruction-biased XML. Possibly, using XSLT.
These options will be explored in the near future.
No action items were discussed
It had been planned to have Sebastien Procureur come and give us an informal presentation on the tracking work he has been doing for CLAS12 at our next offline meeting. However, due to scheduling conflicts it was suggested that his presentation be moved to the tracking meeting. Announcements will be sent out once the scheduling is resolved.