- tactical plan
- ups/upd use
- group activity
- issues and opportunities
notes for meeting¶
To let everyone know what is happening within our group.
To start dealing with the QA issue and efficiently moving forward with our projects.
We do need to start having regular meetings - I propose once a month for the whole
group, for one hour (or less). Individual project meeting are also necessary, at least
for LQCD workflow, mu2e, multicore, and performance database. As soon as the DDS eval
and JDEM R&D project start moving forward, we will need a standard meeting it also.
- discuss job reqs
- discuss the tactical plan status report
- RDI for summer
- new tactical plan - overview
- neutrino computing workshop
- do we or will we need another?
- more memory?
- machine organization
- moving projects into place and off whcdf03
- admin of web server, SVN, redmine, products
- managing virtual machine environment
- performance tool installation needed
Ron talks about his proposed use.
QA - what we need to be doing¶
- need for a brainstorming session,
- keeping it lightweight
- counting visits for answering questions
- categorizing the people making the requests
- counting follow-ups and completions
- use of the redmine system (tracking, wiki, documents, projects)
- how do we make use of such a thing?
- use of wikis
- software process (from time requests to requirements to products to support, DOORs)
- what process do we follow?
- use of SVN (or other management system)
- meeting minutes
- highlights of what each person (of set of people) is doing
- mu2e (off to good start, but how do we get all the work done here?) * storage manager * performance database * monitoring applications - GRID, T0 * JDEM work redirection * jdem QC and LQCD reliability and DDS * LQCD workflow and JDEM DR framework * NOVA * C++ committee * AD and A0
- upcoming workshop of Suzanne
- upcoming Geant 4 work
- need for Ruby/Rails tutorial
issues and opportunities¶
- CRL handoff
- validation framework
- progress on multicore
- workflow specification work
- performance analysis (visit Argonne)
- how to get things done in a timely fashion - project TODO lists
- accelerator modeling performance analysis
- by project
- discuss three levels of framework: control system/DAQ, event processing, workflow
- tracking of scientific data products and processing
- consultation on software engineering, programming, application performance
We did not make it through this whole list. We did not get Ron's presentation of UPS use, or the issues, or the project highlights done.
Walter suggested email as different sort of thing than talking as we might want to consider classifying it separately. Suzanne suggested that it is not really much different for her. Jim said he thought it took more time to response to email.
Somewhat asked how these measurements would be used. In some cases, they will just tell that if we predicted corrected how many times something should occur. In other cases, we will get a baseline, and compare with previous years/months to see if things are better/worse.
Things we could measure concerning consultations:
- complexity - hard, easy, between * answered - no, partial, yes * follow-up - yes, no * email correspondence - yes, no * group - CMS, AD, Neutrino, QCD, etc.
Suzanne asked why not use docdb, JIRA, plone, helpdesk, and other lab supplied facilities.
For one, these are not integrated solutions for project management. Other are
- it is a complete project management application * puts all things together for you in one connected place * connection to stored documents and plot data sources, instead of just rendered output * gives classification and navigation needed for document and source access * Marc gave examples of tex files as source * Jim explained about the source data from plots
Needed with regards to redmine:
- How do we organize documents and projects in redmine? * Should we use subprojects? * can there be cross-linking between projects? * can there be multiple views of information stored there?