Home » Groups » Software and Computing

Offline QA to-do list before Run 14

Updated on Tue, 2016-07-12 10:04. Originally created by genevb on 2013-12-04 14:03.
Based on my presentation at the Run 13 critique meeting:
  • Histogramming codes:
    • (H1) Pull in HFT codes (status: codes started, but never declared ready)
    • (H2) Continuing artifact: many histograms are luminosity dependent (# hits, # tracks, global/primary ratios, flags, DCAs, etc.); at the very least, we should present the luminosity (meaningfully) to the crew, but there’s plenty of room to be even smarter about this
    • (H3) Pile-up vertices in runs without BEMC: runs without BEMC pose a problem for excluding pile-up vertices; really needs improvements to vertex-finder ranking schemes, which is out of our hands
    • (H4) MTD codes? (late: codes inserted into Offline QA post-Run)
  • QA Browser:
    • (Q1) Links to run-specific information need to be updated/added at additional (earlier) steps in looking through available runs (e.g. FastOffline Browser, ESL, RunLog) (status: discussed, but not implemented)
    • (Q2) Combining multiple runs across varying luminosity (not just across fills) introduces confusion; is there a way to discourage this? (status: plan to show luminosities in QA browser started, but not completed; documentation on the topic was improved)
    • (Q3) Manual combine has a memory leak that limits it (can get around it by using AutoCombine, but really should be addressed before next Run)
    • (Q4) Use username in DB queries (no assumptions)
    • (Q5) Fully shake down on new offline web server
  • AutoQA:
    • (A1) Need to think about optimal way to set the threshold “cut” value, more correlation-plotting options
    • (A2) Some histograms really only worth examining in specific triggers
  • Issues / Issue Browser:
    • (I1) Presentation of runs for which issues are active needs improvement [Done!]
    • (I2) Standardization of issue names has additional room for improvement
    • (I3) Reporting of TPC dead channels is still confusing to the crew (anodes vs. RDOs vs. FEEs; what to report and what not to report)
  • Documentation / Instructions:
    • (D1) Clarity (documentation) on how to report issues in non-trigger-typed (general) histograms needs improved
    • (D2) Standardization of issue names has additional room for improvement
    • (D3) Reporting of TPC dead channels is still confusing to the crew (anodes vs. RDOs vs. FEEs; what to report and what not to report) [Documentation was improved]
    • (D4) Training can be developed into a course that can be taken remotely, in advance of shift sign-up
    • (D5) The usual (re)cycling of any documentation from one Run Year to the next, making sure all the links work
  • QA Shift Certification:
    • (S1) Database tables, connected to shift sign-up to warn uncertified persons they need to come to BNL; I will work on this with Dmitry

-Gene

Groups: