• Created by Unknown User (robyn), last modified by Unknown User (abecker) on May 25, 2014

Date

May 23, 2014

Attendees

Goals

  • Blockers, Just Did, Do Next

Discussion Items

Jim Bosch

  • No Blockers,
  • Perry is starting to think about the next sprint. He'll probably be doing more next week.
  • Jim will be traveling next week. He'll try to provide team reports but may not actually know what's going on.

Kian-Tat Lim

  • The SAT is getting started again. He suggests that anyone who had technical design issues to bring them up. So that the SAT can eventually discuss and record them. 

    • Regarding that:  the SAT is currently at around 11 people - which seems rather large - but trying to get all of them in a meeting at the same time is a little difficult. One of the best times seems to be on Tuesday at 11am so we can either keep this Standup meeting very short and have the SAT after it or have the Standup meeting just before or some other day/time. There is another alternative for the SAT meeting but the time currently allocated to the Standup is looking like the best time all around. Once a few more doodle votes come in, the SAT and the Standup can talk more about this conflict.
  • Middleware is going pretty well.
  • The butler has still been on hold.

Andy Becker

(Ask AndyB for any clarification you find necessary. The scribe was unable to regurgitate the whole interaction)

  • Andy is sprinting along on the DCR work. He's been working with the catalog team to go through the entire PhoSim catalog to essentially use a realistic distribution of stars to figure out what DCR is going to actually look like in our data.  He is working towards determing which measurable photometric properties (colors) can be used to estimate the refraction amplitude in a given passband. No blockers on this.
  • Working with Simon, AndyC and Russell to define Simon and Russell's tasks for Su2014.
    • Simon's going to be starting work on fixing up the cameraGeom tasks that were not totally finished: cameraGeomUtils and the Subaru and Decam camera geometry representations.
    • Russell's time may be a bit sparse for Su2014 he's got APO (Apache Point Observatory) work including a TCC (Telescope Control) review in September. So he'll be put in effort as needed.
  • Yusra will be putting together some PSF-matched coadds
    • JimB: What was the goal of the PSF-matched coadds?
    • Andy: To test out image subtraction using PSF-matched coadds
    • JimB: As long as you have images which don't have a lot of distortion, that's ok. If you do have image with distortion, the code is wrong for that.
    • AndyB: We're going to be using the Stripe 82 data as our baseline because there is some actual science which can be gotten out of that.
    • JimB: I think that as long as you're not assuming  in the difference imaging stuff that the PSF on the coadd is what it said it is, you're OK with that.
  • Another thing that Yusra will be working in – we've done forced photometry based on detections in one of the coadds in g or r band. She's going to go ahead and do the union of all things detected in ugriz band coadds. Doing forced photometry in all those positions. To have sensitivity in all things that are really blue or red.
    • JimB: I'm just finishing up a sprint now that had Yusra as one of my team members. Should I not expect her to be one of my team going forward?
    • AndyB: That's a conversation to have with her adviser.
    • JimB: I know she's a grad student so she's not really dedicated to any team.
    • AndyB: We're trying to get her focused on her thesis work. So going forward, she'll probably have less time to work on LSST work. Talk to AndyC for the official info.
  • AndyB is starting to put together a Confluence page to serve as the go-to place for reducing decam data through LSST.
  • K-T: Going back to DCR, a quick question: I read thru your document. At the end you mentioned something that if DCR poses problems for differences, it should pose problems for coadds, too. Do we have different techniques that we can use for them because we have more time or are those still algorithms that need to be worked out.
    • AndyB: I'm envisioning the same algorithm can address both aspects of the problem: the creation of the templates and the mapping of those templates to the science image.
    • K-T: The mapping of the templates to the Deep images, too?
    • AndyB: Absolutely, any of the coadditions: either templates or deep coadds for detection. I'm thinking the same process can ideally be applicable to both situations. The goal of this sprint is to get a baseline code that tries to undo the effects of DCR on an image. So the idea would be that when you're creating a new template for difference imaging, you try to undo DCR and then when you map that to a science image you try to mimic the effects of DCR in the template.
    • K-T: Other people haven't had to worry about this because of their optical design or they haven't tried to do anything at this level?
    • AndyB: They've not tried to do anything at this level. DCR is admittedly a problem in a lot of image subtraction pipelines. And people tend to just ignore the DCR dipoles. This actually held back, for example, the Super MACHO project from getting a lot if its science done. There's one project that did try to do some work on this. Andrew Drake actually tried to do a pixel level correction in the MACHO data for DCR. It worked to some degree but it's a bit tricky since he didn't treat bundles of pixels for PSF. He just did every pixel as DCR affected by itself. It did help the difference imaging residuals.
    • JimB: One more thing: for HSC we're not dealing with DCR due to optical design. And I think that might be true for DES, too.
    • AndyB: What about the optical design?
    • JimB: We have an ADC.

Jacek Becla

  • Jacek was having technical difficulties with his connection. He 'chat' report was:
    • apologies, technical difficulties with hangout audio/video here; no blockers on my side, see qserv mtg notes for all details. right, sorry!!!