Related Material
Available Manpower
| Name | SPs | Note |
|---|---|---|
| Bosch | 43 | Scientist, leader, 0.5 time |
| Gee | 78 | 0.5 time |
| Lupton | 36 | 6 SPs per month |
| MacArthur | 126 | Scientist |
| Price | 63 | Scientist, 0.5 time |
| Swinbank | 87 | Scientist, leader |
| Total | 433 |
Suggestion that it may be more effective to have 80% Bosch & 20% Price rather than 50/50 through this cycle. Need to agree this with the HSC crew.
LDM-240 Goals for FY15
The LDM-240 goals specified for FY15 under the 02C.04 WBS are:
WBS 02C.04.01 (App framework for catalogs)
- Development in support of use cases utilizing catalog data
WBS 02C.04.03 (PSF Estimation)
- Improved PSF estimation and interpolation (Kriging interpolation)
- Integration of wavefront estimation code into the DM stack
Suggested replanning
"Improved PSF estimation and interpolation (Kriging interpolation)" moved to FY17 and replaced with "Estimation of PSF variation across the full focal plane (single solution)" from FY17.
WBS 02C.04.06 (Object Characterization)
Greedy optimizer for multi-epoch data
- Multi-epoch extended source fits
- Micro-optimized MultiFit model evaluation code
Suggested replanning
Astrometric and Photometric Calibration Pipelines (02C.03.07 & 02C.03.08) are currently assigned to UW in FY18+. DRP will likely need a subset of this functionality much sooner than that. Perhaps we can "adopt" part of those tasks, or encourage the UW folks to reprioritize so that the they are addressed sooner.
Summary of W15 Epics
NB "done" totals reflect only stories marked as done; not awarding "partial credit" for in review or similar.
Totals are based on mental arithmetic (so are only approximately correct), calculated on 22 Jan.
WBS 02C.04.01 (App framework for catalogs)
WBS definition
The "Measurement *" epics defined here would be more appropriate under 02C.04.06 (object characterization). Should discuss with Kevin Long/Jeff Kantor if & how to best address that.
Measurement - Aperture Corrections DM-85 - Getting issue details... STATUS
Not meaningfully addressed in W15.
Estimated 31 SPs; planned 31 SPs; done 0 SPs.
This work is well defined & planned.
This is under a DRP WBS, but most of the work is (apparently?) to be done by UW/Krughoff/Owen.
Candidate to roll over to S15.
Measurement - Calibration and Ingest DM-1074 - Getting issue details... STATUS
Aim to have this finished in W15.
Estimated at 20 SPs.
afw::table - finish interface transition DM-1099 - Getting issue details... STATUS
Aim to have this finished in W15.
Estimate 36 SPs. Slightly reduced scope to 23 SPs; remainder stories pushed to DM-1767 which can be scheduled S15.
Measurement - Convert old algorithms DM-1100 - Getting issue details... STATUS
Aim to have this finished in W15.
Estimated 31 SPs; planned 46 SPs. Should be done in W15; a few remainders pushed to DM-1769 which can be scheduled S15.
Measurement - Finish framework overhaul DM-1101 - Getting issue details... STATUS
Aim to have this finished in W15.
Estimated 51 SPs; planned 55 SPs; done 45 SPs. Should be done in W15. Remainders pushed to DM-1769 which can be scheduled S15.
afw - Footprint improvements DM-1107 - Getting issue details... STATUS
Unlikely to be finished in W15.
Estimated 30 SPs; planned 53; done 15.
Measurement - Design MultiFit Plugin Framework DM-1109 - Getting issue details... STATUS
Candidate to roll over to S15.
Estimated 20 SPs; unplanned; done 0.
NB, implementation of this plugin framework to be scheduled in S15 after design.
WBS 02C.04.03 (PSF Estimation)
No work planned in this cycle.
WBS 02C.04.06 (Object Characterization)
Work on ideas for the deblender DM-1382 - Getting issue details... STATUS
This epic needs refocus and refactoring.
Not estimated; planned 140 SPs; done 0.
We've shut down (Won't Fix) the two might 140 SP stories, and used some development work done on HSC to close other stories out. That work will be ported to LSST probably in S15. The story for that porting is nominally DM-533, but in practice that will need to be expanded.
Full list of HSC deblender issues to be ported to LSST is at https://hsc-jira.astro.princeton.edu/jira/issues/?filter=11603.
There will also be new LSST deblender development on top of the above.
Other outstanding work
There are some Getting issues... in JIRA with no defined epic. They cannot therefore be scheduled for EVM. Most of these are relatively minor bug fixes or feature requests; those with SP estimates have no more than 4 SPs per issue. (That said, 35 issues at 2 SPs per issue is still 70 SPs, which is a major epic's worth of work).
We should consider if we can either add some of these to existing epics or combine them to create new ones.
HSC work which will be carried out in S15 which we can port to LSST
?
Outline S15 epics
Rough priority listing:
- Finish measurement framework overhaul
- Deblender
- Galaxy fitting
- Multi-Fit plugin framework
- PSF estimation
- Photometric/astrometric calibration pipelines
We'll likely have to schedule a ~20 SP epic for toolchain enhancements (GitHub migration, etc).
WBS 02C.04.01 (App framework for catalogs)
| Epic | SP estimate | Likely assignee(s) | Note |
|---|---|---|---|
| DM-85 | 31 | Bosch + UW | Rolling over from W15 |
| DM-1767 | 27 | Bosch, Gee | afw::table improvements; roll-over from DM-1099 + some new work |
| DM-1107 | 23 | Bosch, Swinbank | Footprint improvements. Rolls over from W15. |
WBS 02C.04.03 (PSF estimation)
| Epic | SP estimate | Likely assignees(s) | Note |
|---|---|---|---|
| (not yet created) | ? | Swinbank | PSF estimation |
| (not yet created) | ? | Swinbank | Wavefront estimation |
WBS 02C.04.06 (Object Characterization)
| Epic | SP estimate | Likely assignee(s) | Note |
|---|---|---|---|
| (not yet created) | 60 | MacArthur, Price | Porting from HSC and other deblender infrastructure |
| (not yet created) | ? | MacArthur | LSST deblender development |
| DM-1768/69 | 26 | Gee, Bosch | Rolled over stores from the W15 measurement framework reworking + some new work. DM-1768 contains low priority miscellanea and doesn't need to be scheduled this cycle, but DM-1769 should be. |
| DM-1069 | ? | Gee(?) | Measurement unit tests. Depends on DM-1102 (SQuaRE). |
| DM-1108 | 76 | Gee (with support from Bosch, Krughoff) | Galaxy Fitting - Shear Precision Experiments |
| DM-1111 | ? | New C++ developer if available | Galaxy Fitting - Micro Optimization. Not essential for our development goals to address this epic in this cycle, but may be necessary to inform hardware choices. Check with Mario. |
| DM-1110 | 60 | Bosch | Galaxy Fitting - Sampling Algorithm Plugin |
| DM-1109 | 20 | Bosch | Design MultiFit plugin framework. Rolled over from W15. Depends on updated Butler (or we stub in our own temporary Butler-alike hack, which would be less than ideal). |
| (not yet created) | 50 | Gee | Implement MultiFit plugin framework. |
Total SP allocations
Based on outline tasks above; simply dividing equally among likely assignees for now.
| Assignee | SP |
|---|---|
| Bosch | |
| Gee | |
| Lupton | |
| MacArthur | |
| Price | |
| Swinbank | |
| UW/Krughoff/Owen |
Dependencies on other groups
SQuaRE: DM-1102, required for our DM-1069.
UW: DM-85. Astrometric & photometric self-calibration (which specific issues do these block?).
Middleware/Butler: DM-1109, multifit, etc.
To Do
- Agree plan for DM-85 with UW group
- Agree plan for DM-1102 with SQuaRE
- Agree plan for astrometry/photometry pipelines with UW group
- Discuss shifting measurement epics to 02C.04.06 with Jeff/Kevin
- What work will be addressed by HSC during S15 that we can claim partial credit for?
- Determine status of wavefront estimation code & consider an epic to integrate that with DM stack (02C.04.03)
- Agree allocation of Price/Bosch with HSC
- Estimate SPs for epics remaining above
- Question for Jeff/Kevin: For epics which we have only started to address in W15, should we close them down and create new epics for S15, or just carry the same epic over?
- Discuss timescale on which DM-1111 is required to inform hardware choices.