Shape Model Review 1 Feb 2016 Suggested resolution of liens as of 10 Feb 2016 Initials of lien submitter are in [ ]. Liens have been paraphrased and combined as appropriate. None of the liens on the SBN side will be addressed until after the major Rosetta review next week. 1. Dataset organization is confusing to someone not already familiar with the various kinds of models [CA]. *** A’Hearn has drafted and is now editing (noting comments from Acton) an overview document to outline the directory structure and the various types of models included. This will need to be updated periodically as additional models are added to the dataset. This will go in the document directory but will be pointed to by the aareadme.txt file at the dataset level. Action A’Hearn with assistance from Farnham and Barnes. 2. Need document on SPC vs. MSPCD (both listed in SPC directory) [CA, RK] *** A document needs to clarify why various models were (and will be) created, how they are related, and when to use each one. Laurent’s document addresses SPC vs. MSPCD but not the other types (ESA, DLR). Action SBN, Farnham with assistance from A’Hearn? TBD 3. Directory structure/names is misleading. [CA] *** The names at the top level may be misleading, since all the models are tri-plate. Item 1 above partially addresses this and the first sub-level name DSK may be changed to SPICE_DSK. Action Farnham and Barnes 4. Describe the imaging periods in the aareadme.txt and the index table (and other comments on index table. [BG, RK] *** A separate scientific index will be created. The basic index table required by PDS is generated by a script to get the paths straight and as such is best kept simple in content. The time span of observations used for each model will be in the scientific index table. For each sub-directory and/or product we will consider adding the range of observation times to the aareadme.txt file. Other useful fields, such as instrument name(s) & number of plates, will be in this index. Action Farnham and Barnes to organize, Barnes to implement, A’Hearn to assist where appropriate in locating the details. 5. Missing references in reference.cat [RK] *** Action SBN - only references in catalog files are in REF.CAT, not references within other objects. 6. DSK format of ESA (RMOC) model has errors. [BS] *** We will substitute the NAIF model. Action Farnham. *** Labels should be coordinated across ESA and OSI-NAC models. DATA_TYPE needs to be fixed (discussed here at UM). Action Barnes to manually fix ESA label. Action PSA to get relevant information to data providers for use in future deliveries from pipelines. 7. What is PDS standard for shape models? [RK] *** Words “PDS standard” regarding the “tri-plate” models should be changed to reflect that this is a “standard” developed by SBN long before SPICE dealt with shape models and is aimed at use in a wide variety of non-SPICE tools. DSK will become a PDS standard for use exclusively in SPICE (no other software knows how to read it) when there is a public release (as opposed to beta releases available only on special request) of the SPICE toolkit (anticipated spring 2016). Action Farnham. 8. EAICD problem [TB, MA] *** Parent document is incomplete abd will continue to change. Suggest for now including only annexes, where available, in relevant places. *** Annex for ESA model is inadequate for explaining how the model was derived and there is no published paper to reference for the information. Action PSA (O’Rourke?) 9. ESA filenames do not follow OSI filename structure [RK] *** This would be desirable, particularly if there will be future deliveries of later models and this can be incorporated in pipelines. We can change this at SBN but it has to feed back into future deliveries so it is better handled at PSA. Action Barthelemy. 10. Offset (“flexure” between head and body) in ESA vs. OSI [BG] *** Needs to be described in CONFIDENCE_LEVEL note of dataset.cat. Should probably be also mentioned in overview document comparing the models. Action Barnes for dataset.cat, Farnham or A’Hearn for overview comparison document. 11. NAVCAM model not available in multiple resolutions. [RK] *** This is important for optimizing the speed of calculations in projects that do not require the full resolution. If possible, action RMOC & Barthelemy? 12. NAVCAM model does not have .png images to help users check that they are using the model correctly. [MA, & mentioned by others during review] *** Needed for users to be sure they are using the model correctly. Action RMOC & Barthelemy? 12. Many editorial points in various presentations. [var] *** Action Barnes 13. Some files do not meet 27.3 [DH] *** This needs to propagate back to pipelines of data providers for future models. Action Barnes 14. DATASET_ID is wrong in dataset.cat [BS & DH] and in voldesc.cat [BS] *** action Barnes 14. Various other PDS-PSA inconsistencies a. PRODUCT_ID missing in CITATIONS.LBL. Action Barnes b. PRODUCT_ID missing in CHECKSUM.LBL. Checksum should be regenerated at each stage of data transfer and PSA has deleted the checksum file for other instrument datasets. We suggest that PSA simply delete our CHECKSUM file. c. ARCHIVE_STATUS missing in DATA_SET_INFORMATION. Since the value of this keyword should be changed several times during the archiving process (and as recognized in PDS4 should not be included in the dataset itself at all), the value will be wrong unless it is manually changed at each stage of the archiving. As PSA knows, our solution has been to drop the keyword, lest it end up with the wrong value in the final archive. If PSA can tell PVV to ignore it, that is best. If not, we suggest inserting the keyword in the PSA copy. This is not a difference that will matter to a scientific user. d. SOFTWARE.CAT is missing. This is an issue on which identity with the data between PSA and PDS is meaningless. When we get datasets from PSA that include SOFTWARE.CAT, we do NOT put that in the central catalog at EN. If PSA can insert a dummy SOFTWARE.CAT file, that is the easiest approach. No scientific user will be affected. Action PSA. e. Keywords missing from INDEX.LBL. These keywords will be in our scientific index (item #1 above) - action Barnes. For the PDS required index, it is our understanding that DVAL forces regeneration of this file anyway, so there is no guarantee of a match between PDS and PSA even if we did include those keywords. No action. f. Missing keywords in data LBL files. As noted by DH, these will be dummy keywords set to N/A. And this then is something that will have to be fed back into the pipelines of all the shape model providers for future models. We would prefer to have PSA insert these only on the PSA side (noting the large number of files to which it applies retroactively). f. Pointer format () vs {}. This is a case in which either form is valid, albeit with different implications (ordered sets vs not-ordered sets). However, the PSA validator flags {} as wrong, while the NASA Ames validator (vtool) flags () as wrong. In fact, both validators are “wrong” according to the actual reading of the standard, since the validator does not know whether the set is ordered or not, and at least in this case as is probably true for most cases, the difference in interpreation (ordered vs. unordered) is irrelevant. If PVV can be told to ignore it, that is best. If PSA wants to change their version to (), that is fine with us since it will not affect any scientific users. Action PSA.