Child pages
  • October 22nd 2014
Skip to end of metadata
Go to start of metadata
Attendees: Chris, Jim, Lynne, Scott, ajc, Kem, Cathy, Heather, Brittany, Veljko, Zeljko, Simon, Peter, Darko, Heather
 
Call in number: 866 330 1200
Code: 518 2287#

 Goto Meeting  ID  821-177-359
Agenda for this week:
  • Tier 1 simulations: what is in them and what do the results look like (Kem)
    • Tier 1 Overview
    • Tier 1 runs complete except (3) with 3 visits
    • Design area and design number of visits take about 80% of the time in a 10 year run
      • Simulator allows overflow of the  number visits by doing the more of the same (i.e. more depth)
      • SRD numbers are completed in 8 years (can just ask for more visits which is called boosting)
      • lots of filter changes for the final year
        • fixed if you restart the proposals when completed (or allow the run to go for 15 years and that the first 10)
      • Too many visits for the NES (12% but we probably want about 2-3%)    
      • Zeljko - will have a solar system metric soon to determine how much time we need in the NES
    • 2.5-3 days of CPU time for 10 years of a run
      • Chris - can you run 10 jobs at once (yes)
      • Chris - can you run year 1 and year 10 at the same time (no)
    • Code is non-deterministic code. It is deterministic for a few hundred days (fields have identical weighting but different iterations of the code selects different fields - even if we say select the lowest value fields
    • Are they ready to release to the community? No
      • release v3.2 of opsim by the end of week: 
        • ra limit fixed for rolling cadences, restart ability for WFD to give a uniform end of survey
      • rerun Tier 1 at this point
  • Update on proposal to transition to Github
    • likely it will happen - DM will provide a transition plan
  • Ethnography: questions for Brittany about her process
    • Some worries about how having someone listening may impact how we work. 
    • Some worries about the breadth of the consent form
      •  Brittany - doesnt use video, audio is interview only, you can specify when you don't want something recorded, more interested in looking at the social interactions (emerging phenomena) than individual statements, interested in working with group rather than just observing, approvals are on a case-by-case basis and you can opt out anytime or for any part of a conversation
      • Brittany will send round a paper that is an example of the type of work she does

 

  • Ran out of time at this point
  • Status  (requirements documents, MAF delivery of SSTAR, solar system metric, IGM code)
  • Best practices for bringing new people on board (Michael)
    • what do we need to teach - github, code reviews, unit tests, build system and installing, maf?, see where his python stands, starter project (YES)

 

  • No labels