Present: Joanne Bogart, Toby Burnett, Richard Dubois, Heather Kelly, Sean Robinson, Leon Rochester, Alex Schlessinger, Tracy Usher, Karl Young
Gleam To-do: (Toby) See his list, which is at least close, if perhaps not identical to, the set of items to be completed for the May Release.
One crucial item is the third bullet, the ability to write digis to a ROOT file. (Heather) The May 2nd date for this will have to be pushed a week or so. Currently Gleam can write hits (both integrating and position), but not yet digis. In fact the digi classes, at least for TKR, are still not quite firm; they should be stable shortly after tomorrow's Tracker Data Structures Review.
There are still some lingering problems with MeV/GeV. Also we need to do something about the source library. One solution would be to add a units attribute to the XML description (which could default to GeV, minimizing changes to existing files) to make explicit what units are used and allow code to do conversions as needed.
Richard confirms that CAL is cognizant of the need to incorporate noise. Either Sasha or Berrie will be handling it.
CalRecon is about ready to tag.
AcdRecon has further to go. For one thing, first need stable ACD digi classes!
(Toby) Would like to test existing abilities to output hits. He will need latest versions of RootIo and mcRootData packages. Heather will tag if necessary.
Tracker Data Structures Review: (Leon) Takes place tomorrow morning (2 May) at 7:30. Could run over and partially overlap the general software meeting. Heather and Norm Graf will be the reviewers. The maximal scope is all Recon classes; minimal would be just TDS classes. Actual scope will probably fall somewhere between. There is a tendency to design on the fly (as code using the classes is being written), which changes the character of the review process.
(Richard) Probably should have some similar oversight for CAL Recon classes.
Nightly builds: (Karl) They're working again, using glastpack. All changed packages are built against most recent (compatible) tagged versions of other packages. Would like to institute another build mode in which related HEAD versions would be built against each other.
It is still the case that not all error messages are extracted in the primary build results page; given the form of the build output, there is no simple or even not-so-simple way to do this. However, the full build output can be seen by clicking on the STDERR link.
glastpack: (Alex) It has gained some new functionality since last week: can now use it to build and run apps. He is starting work on a web interface. He and Janani are working to incorporate instructions for glastpack into the How-to web pages. People are encouraged to try it and report experiences. Toby requests that the response to incorrect usage be more specific to give the user some idea of what's wrong. Right now the response is just to print the complete usage help.
Future VCMT: (Toby) He has an alpha version of a new VCMT using the C# language and .net. See this page for a screen shot, status and some pros and cons.
It's not clear how this fits into the overall scheme of multi-platform support. Alex confirmed that non-Windows-compatible aspects of glastpack, if any, are not essential and could be modified in short order.
Calibration status: (Joanne) Madhup now has a workable development environment. (Thanks, Alex, for installing the client MySQL includes and binaries in the standard external library spot for SLAC Linux machines.) He will write a simple application at least partly in c which accesses the database using the standard MySQL client library (itself in c). He is also investigating the use of MySQL++, a contributed library providing a C++ API. Joanne is already on version 2 of an XML bad strips format (dtd) without having yet produced a single XML document. For this version hot and dead strips will be in separate documents.
Geometry: (Joanne) Just finished committing modifications to xmlGeoDbs for tracker MCM boards. Will tag after implementing tracker walls. Meanwhile discovered some disconcerting behavior in handling of materials which needs to be understood.
More student news: (Richard) Janani's work on the How-to pages can be seen in the Frontpage developement area, http://www-glast.slac.stanford.edu/software/devWeb/CodeHowTo/
Deepu is investigating the ROOT-Apache module, aka Carrot. It is up and running on centaurusa and she has already used it successfully.
J. Bogart Last Modified: 04-Aug-2004 15:41:53 -0700