Jeffrey Carlin has identified a number of requirements that are referenced in the DMSR but that do not exist in the DMSR (or anywhere). This is reported in - DM-11029Getting issue details... STATUS
The DM-SST will review these; if they are relevant, we will propose via LCR to to include them in the DMSR, if not, the dangling references to them should just be removed.
Criteria for evaluation
- Verifiability - can the requirement be verified by DM?
- Feasibility - is it feasible to verify a requirement
- Useful - is it useful to add these at this late stage?
Should we make the DMSR a stand alone document?
DMSR 'dangling requirements'
Column | Specification | Retain/Modify/Delete | Comments | Actions | Status |
---|---|---|---|---|---|
DMS-REQ-0003: Create and Maintain Science Data Archive | The LSST project shall create and manage an archive of all its public data products and the raw data necessary to reproduce them. In addition, the archive shall contain all necessary engineering and calibration data for the full understanding of the performance and operation of the Observatory. |
| Requirements deriving from this:
Derived from Requirements:
Which OSS requirement was this meant to shadow? OSS-REQ-0167 |
| |
DMS-REQ-0005: Produce Data Releases | All Level 2 Data Products shall be made public as part of periodic coherent Data Releases |
| Requirements deriving from this:
Which OSS requirement was this meant to shadow? | DMS-REQ-0005 looks to derive from OSS-REQ-0134: 3.1.5.3.1.1 Level 2 Data Product Availability ID: OSS-REQ-0134 Specification: All Level 2 Data Products shall be made public as part of a Data Release, with releases coming at least once every time DRT1, and as soon as possible following the completion of data release processing as is consistent with verifying the applicable data quality requirements | Removed as part of LCR-962 DONE |
DMS-REQ-0007: Pipeline Infrastructure | The DMS shall provide Processing, Storage, and Network resources capable of executing the DMS Data Release pipelines over all pre-existing survey data in a time no greater than drProcessingPeriod. |
| Has been put back as DMS-REQ-0163 - it is a near-exact duplicate of the original -0007 text. | DONE | |
DMS-REQ-0009 |
| DMS-REQ-0009 is documented as being derived from DMS-REQ-0007 but that relationship makes no sense. | Remove remove dangling requirement DMS-REQ-0007 from 3.2.1 Simulated Data ID: DMS-REQ-0009 (Priority: 1b) | TODO | |
DMS-REQ-0011: Produce Difference Sources | The DMS shall produce Difference Sources from the analysis of Difference Exposures. |
| Requirements deriving from this:
LPG: Related is DMS-REQ-0269: The DMS shall construct a catalog of all Sources detected on Difference Exposures with SNR > transSNR. Looks like this requirements was meant to shadow 3.1.5.2.1.4 Catalogs (Level 1)
| Remove dangling requirement DMS-REQ-0011 from 1.3.3 Difference Exposures ID: DMS-REQ-0010 (Priority: 1b) Flowdown from intended OSS requirement is captured in DMS-REQ-0269: | TODO |
DMS-REQ-0040: Enable BAO Analysis | The DMS shall provide the data products enabling the measurement of the effects of Baryon Acoustic Oscillations (BAO) in the detected galaxies. |
| Requirements deriving from this:
This requirement does not state any higher level requirement that it derives from This was probably an attempt to take on more that what was in the SRD. The idea was that somebody should sign off to produce the inputs that are need ed - i.e provide photometric redshifts See comment 4b) below | Remove dangling requirement DMS-REQ-0040 from 1.4.2.1 Provide Photometric Redshifts of Galaxies ID: DMS-REQ-0046 (Priority: 2) | TODO |
DMS-REQ-0041: Measure Intrinsic Ellipticies of Small Galaxies | The DMS shall measure the intrinsic ellipticities of small galaxies. The maximum galaxy size for this requirement is 10 arcsec. The measurement error goal for this is TBD. |
| Requirements deriving from this:
Related to OSS-REQ-0161, which was recently rewritten in LCR-1405 to not specify requirements on the measurement of galaxy ellipticity DMS-REQ-0047 also derives from 3 OSS REQs:
L: Requirement to provide PSF is not only driven by the need to measure ellipticity of galaxies. See comment 4b) below | Remove dangling requirement DMS-REQ-0041 from 1.4.12 Provide PSF for Coadded Images ID: DMS-REQ-0047 (Priority: 1b) | TODO |
DMS-REQ-0055: Correct for Camera Bias Structure | GPDF: Possibly meant as a derived requirement from another requirement in the DMSR |
| Requirements deriving from this:
Derived from Requirements:
| Remove dangling requirement DMS-REQ-0055 from 1.5.4 Bias Residual Image ID: DMS-REQ-0060 (Priority: 1a) | TODO |
DMS-REQ-0056: Correct for Camera Crosstalk | GPDF: Possibly meant as a derived requirement from another requirement in the DMSR |
| Requirements deriving from this
Derived from Requirements:
| DONE | |
DMS-REQ-0057: Correct for Detector Fringing | GPDF: Possibly meant as a derived requirement from another requirement in the DMSR |
| Requirements deriving from this
Appears to be a duplicate of
| Remove dangling requirement DMS-REQ-0057 from 1.5.7 Monochromatic Flatfield Data Cube ID: DMS-REQ-0063 (Priority: 1b) | TODO |
DMS-REQ-0058: Correct for Instrument Sensitivity Variation | GPDF: Possibly meant as a derived requirement from another requirement in the DMSR |
| Requirements deriving from this
Related to?
| Remove dangling requirement DMS-REQ-0058 from 1.5.3 Bad Pixel Map ID: DMS-REQ-0059 (Priority: 1a) | TODO |
DMS-REQ-0066: Keep Exposure Archive | The DMS shall provide an archive of all Raw Exposures taken by the LSST. All other exposure types produced by the DMS (Processed Science Exposures, Calibration Exposures, Coadded Exposures) shall either be archived, or a method shall be provided to recreate them on demand. The Archive shall be designed to be capable of managing the data for the full duration of the survey and associated data processing. |
| Requirements deriving from this:
Derived from Requirements:
| Readd to DMSR | |
DMS-REQ-0076: Keep Science Data Archive | The DMS shall provide a science data archive containing catalogs of all detected sources and objects and associated metadata that meet data quality thresholds for retention. |
| Requirements deriving from this:
Related also to OSS REQ: OSS-REQ-0004: The Archive Facility. Specification: The LSST shall provide an "Archive Facility" to host the following functions:
Appears to be a duplicate of
Discussion: The Archive Facility includes the personnel offices, computer equipment, and other specialized infrastructure necessary to safely execute all the functions of the listed Centers and to secure all LSST assets located at the Archive. | Replace dangling requirement DMS-REQ-0076 with DMS-REQ-0003 in
| TODO |
DMS-REQ-0080:Difference Sources Available within 24 hours | Detected sources and associated metadata shall be available for public access in the DMS science data archive within 24 hours of their generation by the DMS. |
| Requirements deriving from this
DMS-REQ-0080 focuses on the latency whereas DMS-REQ-0033 is about simply providing software to detect sources. Appears to be a duplicate of
Propose to delete and replace derived from requirement under DMS-REQ-0033 with DMS-REQ-0004 | Replace dangling requirement DMS-REQ-0080 with DMS-REQ-0004 in
| TODO |
DMS-REQ-0081: Produce Object Catalog | The LSST DMS shall produce a catalog of all objects detected by the Detection Pipeline in at least two Difference Exposures from a single Visit. |
| Requirements deriving from this
Appears to be a duplicate of
| Replace dangling requirement DMS-REQ-0081 with DMS-REQ-0275 in DMS-REQ-0034 | TODO |
DMS-REQ-0086: Produce Orbit Catalog | The LSST DMS shall produce a catalog of all moving objects detected by the Moving Object Pipeline (MOP) |
| Requirements deriving from this
Appears to be a duplicate of
| Replace dangling requirement DMS-REQ-0086 with DMS-REQ-0273 in DMS-REQ-0089 | TODO |
DMS-REQ-0090: Generate Alerts |
| Requirements deriving from this
Appears to be a covered by DMS-REQ-0274: 3 Alert Content ID. The DMS shall create an Alert for each detected DIASource, to be broadcast using community protocols, with content that includes: a unique Alert ID, the Level-1 database ID, the DIASource record that triggered the alert, the DIAObject (or SSObject) record, all previous DIASource records corresponding to the object (if any), and cut-outs of images (from both the template image and the difference image) of sufficient areal coverage to identify the DIASource and its immediate surroundings. These cutouts should include WCS, PSF, variance and mask information. The Alert should also include program and/or scheduler metadata. No requirement derived from DMS-REQ-0274. Propose to delete DMS-REQ-0090 and use DMS-REQ-0274. Discussion: Would it be clearer to rename DMS-REQ-0274 to 'Generate Alert Packet' or 'Generate Alert Packet' ? | Replace dangling requirement DMS-REQ-0090 with DMS-REQ-0274 in
| TODO | |
DMS-REQ-0092: Alert Attributes | For each alert, DMS shall store the pipeline parameters used to derive it, and a set of metadata attributes including at least:
|
| Requirements deriving from this
I see no duplicate or coverage in any other requirement covering Alert provenance LPG/EB: : DMS-REQ-0094 Keep historical alert archive will archive all the alerts 'as-sent'. This covers the postage stamp and associated information & confidence parts. We do not classify alerts. The pipeline parameters should be part of an eventual provenance solution | Remove DMS-REQ-0092 from DMS-REQ-0094 | TODO |
DMS-REQ-0104: Produce Co-Added Exposures | The DMS shall produce Co-Added Exposures combining the set of exposures taken of the same region of the sky that meet a user-specified set of quality and other selection criteria. These may include site conditions (seeing, sky brightness), telescope conditions (wavefront quality), PSF shape and spatial variability, date of exposure, or filter used. Each Exposure contributing to a Co-Added Exposure shall be projected onto a common reference frame using a projection selected from a set of options by a Policy. [ACTION: harmonize with co-add pipeline requirement, enumerate the types of coadds required] |
| Requirements deriving from this
Appears to derive from OSS-REQ-0136: Co-added Exposures Seems to be (partially) covered by DMS-REQ-0278: Coadd Image Method Constraints There are several type-specific coadd requirements,
all that derive from OSS-REQ-0136: Co-added Exposures. Suspect that DMS-REQ-0104 meant to be one requirement deriving from OSS-REQ-0136 with the type-specific coadd requirements in the DMSR deriving from DMS-REQ-0104 Following the reasoning of 4b) in the comments below, I propose to re-add this requirement to make the DMSR more standalone. | Readd DMS-REQ-0104: Produce Co-Added Exposures. Add enumeration of type-specific coadd requirements listed in DMSR already. Change type-specific coadd requirements "derive from" to DMS-REQ-0104 from OSS-REQ-0136 | TODO |
DMS-REQ-0116: Extended Object Shape Parameters | For extended objects only, the summary metadata attributes must also include shape parameters. Weak lensing studies require the image moments (or equivalent) through N=2. For weak lensing extended objects of size greater than TBD (10 arcsec?) do not require shape parameters to be measured |
| Requirements deriving from this
There is no mention of extended objects in the OSS A related requirement is:
Related: (LCR-1405): OSS-REQ-0161: Deep Detection and Measurement Quality. The Object catalog shall include shape measurements that permit the recovery of constant applied shears between 0.01 and 0.05 from an ensemble of isolated galaxies, with a multiplicative bias less than 0.001 and additive biases less than 0.0001.
OSS-REQ-0161 is not anywhere flowed down to the DMSR. Should DMS-REQ-0116 be associated with OSS-REQ-0161? As for DMS-REQ-0040 above This was probably an attempt to take on more that what was in the SRD. See comment 4b) below | Remove dangling requirement DMS-REQ-0116 from DMS-REQ-0070 | TODO |
9 Comments
Leanne Guy
Gregory Dubois-Felsmann Thinks that most of these were supposed to be in the DMSR but there was not enough time to include them before the FDR.
Gregory Dubois-Felsmann
To expand on that, because more of the story came back to me in the course of meeting with Leanne Guy just now:
<copy>
relationship, but that was not available in a useful way in EA at the time, so as an interim decision it was decided not to duplicate them "by hand" in the DMSR at all (thereby making the DMSR not a truly standalone document, but only usable together with the OSS). This was not seen as a permanent decision, but we never actually went back to fix this up. In those cases where quasi-OSS-duplicates, like DMS-REQ-0003, had been present in the DMSR, we decided to remove them and only put them back, eventually, in the "correct" way using the<copy>
relationship. This is now possible in MagicDraw. It would be extremely useful to complete this work now so that the DMSR can finally be a truly standalone document; that will greatly facilitate verification.I volunteered to look at the "4a" class of requirements and see which OSS requirements could be used in the DMSR via
<copy>
to replace them.Leanne Guy and I will continue looking at this next week.
Michael Wood-Vasey
Leanne Guy I read through the above and Gregory Dubois-Felsmann 's explanation and plan make sense to me.
I understand that the plan is:
4a - Copy from OSS: Gregory Dubois-Felsmann will provide recommendations and ask Systems Engineering to implement.
4b - Detritrus: Gregory Dubois-Felsmann will tell System Engineering to drop these
4c - ISR+Camera: Michael Wood-Vasey and Robert Luptoncan take responsibility for these, but with a low priority. These requirements will be hard to write down in useful way.
Gregory Dubois-Felsmann
This has become relevant to RFC-709 - Getting issue details... STATUS this week. Picking up the ball again.
Gregory Dubois-Felsmann
I'm proposing on RFC-709 - Getting issue details... STATUS that modified versions of DMS-REQ-0003 and DMS-REQ-0066 be restored to the DMSR to support a more logical document flow and improve the flowdown of the LSR-REQ-0049 requirement for on-demand re-creation of un-archived data products.
Leanne Guy
This was done
Leanne Guy
Gregory Dubois-Felsmann I have completed a concrete proposal based on the comments above. Could you review please.
Leanne Guy
Jim Bosch Could you review proposal for DMS-REQ-0104: Produce Co-Added Exposures and comment please
Jim Bosch
The proposed action for DMS-REQ-0104 sounds good to me; I like that it helps a bit with the interpretation of the word "user" in its spec, which otherwise makes it sound like science users are telling us the criteria that sets what goes into our coadds; with those other existing requirements as dependencies, I'm comfortable with that "user" referring to whoever is configuring the pipeline for one of those coadd types (i.e. actually a DM developer).
That said, I am a bit confused by the reference in the comment in the table to Gregory Dubois-Felsmann (4b) bullet, as that seems to recommend the opposite (dropping the requirement) of the action being proposed for this one.