DM Sessions 3 and 4: Notes by JCM and Alberto Micol Characterization Some discussion of STC and Characterization interaction. Asked FB to give some more detailed examples of Char and of his version of Footprint. Asked AR to modify these examples for STC compliance but in the spirit of the CDS approach. It was pointed out that the use of utype namespaces in VOTABLE required a namespace declaration method that doesn't exist yet. is used but there's nowhere in the document that defines the prefix "cha:". Several options: -Add utypens attribute to VOTABLE e.g. RESOURCE: or something like that. -Globally define IVOA data models; use Characterization.Location... instead of cha:... -Globally define prefixes for data models in an external document, declare via xlink in votable somehow. We need to discuss this further. Igor presented a new access protocol using Characterization utypes and ucds for a generic method. Commonalities of DAL protocols -> use characterisation dm Example of DAL access based on char utypes were given. -DT: similar to DAL plans: ADQL and General Dataset queries. -BT: against ADQL which requires yet a new translation layer. ORM layer was suggested. -AR: Why using utype http-post parameters instead of posting an XML document? What fields should be mandatory in Characterization? Options were offered ranging from only Location/Bounds being mandatory (suggested by Jonathan) to Support being also mandatory (the latter championed by Alberto especially). An initial floor vote resulted in the more restrictive option of Support being favored 3-1 with most of the room abstaining. The chair cajoled several people into voting until he got a 4-3 result which he then claimed was close enough, in view of the massive abstaining vote of ~30, to let him overrule the vote and say that there would be two levels of compliance: level 1 with Bounds only and level 2 with Support. It was also stated that only spatial, time, spectral axes were mandatory. It was pointed out that if other axes were present, it would be easier for the schema point of view to make their location/bounds also be mandatory-if-axis-was-present (so that then minOccurs=1 in the schema could be used to enforce this.) Chair did not rule on this. => In the Registry it shall be possible to distinguish among level 1, or level 2 characterisation. Discussion from now on should be public on the dm mailing list, not among authors. Dubernet, Osuna, Guainazzi, Salgado, Roueff on Atomic & Molecular Lines DM: MLD reported on the status of the model. -- Outline: AML DM Version 0.5 (30/1/2006) Goal: achieve PR in 4 months (max). Useful for ETL studies and calibration Line: (Species), (Process, Environment), (Level, QuantumState, QuantumNumber) as main objects. Need to add Accuracy, Quality, Versioning, Documentation. Implementation on Molecular DBB (CDMS, JPL) by VO-Paris in advanced state. Need more general DM and Access Protocols for A & M Physics. Comments: Line Intensity (moved to Environment?) and Title needs some further discussion. Need to add astronomical object (esp for case of unid lines) -IKamp: Huge atomic data providers are going to accept this? Answer: collaboration on going with various groups (nist, united nations, etc.) Not many molecular physics archives around, and Dubernet is doing the conversion for those. -JCM: Any fundamental changes required? No answer from the audience. -FGenova: Interoperability across different disciplines is to be stressed, and is an important result of this IVOA work. -IK: Any activity onto molecular reaction data model? Answer: collisionalData will include this, but not yet there in this version. Leave till later. -AW: how to deal with lists of unidentified lines? Answer: Assignment field will be set to "U", -AW: Many different ways to describe lines (aliases), is that taken care of? Answer: Yes, the title of the line is free text but the model specifies species and levels. Arnold on STC: relevant comments Transformations are not part of the coordinate systems, and were implemented by David Berry (Starlink AST library); current status unclear. Applications of STC: Resource profile, Observation description, Search location specification, Catalog entry location. XML changes since Madrid: Single schema! (from stc: crd: reg: to stc:), XLINK referencing model for IVOA, Accommodate code generators (Java C#, C++). Support for: VOEvent, Footprint service, Registry coverage, SED, generic coordinates, ADQL. STC clients may (and do) decide to implement a subset of capabilities. xlink reference example: Support developments: VOEvent (compromise, added support for orbital parameters), Footprint/Region web services (Budavari). STC, in VOEvent context, usage limited to a dozen coordinate systems: Time: UTC, TT, TDB Position: FK5, ICRS, solar under consideration Reference position: geocentre, ... STATUS: v1.30 schema published. Works with JHU region web service. Incorporated in VOEvent, Registry, SED draft. PLANS: update PR and resubmit, then sort out solar coord sys, develop standard components Xlink library, Develop STC Java library. Comments: -AM: Registry for "standard" lists of IDs? Answer: A number of standard lists of coordinate systems, observatory locations etc. will be registered and made available; requires interaction with IVOA Registry WG. Mark Taylor: Interoperable implementation of the STC is available? Answer: Usage of Coverage in Registry, use of element of STC in VOEvent, JHU Footprint web service. Mark expressed concern that these implementations were insufficient and incomplete, and could lead to discovery of problems later on when the scheme is fully used. AR and JCM felt that the implementations were sufficient to give confidence in the basic model and were comparable to implementations accepted for other IVOA standards. -AM: Does the document provide a list of UTYPES? JCM took an action to provide a mapping to UTYPE within one month. -AM: Observation description example? Answer: see web page for more examples. -FG: Do you feel there are still points where compromise not yet reached? Answer: We have them all at this point, though not yet complete (eg solar system). Healpix type of coord sys still missing. We do need dialog with Characterisation. Answer by JCM: much more mature, with compromises already taken, other differences do not require changes to STC. -FB/FG: Not for Characterisation either! -JCM: Right, neither should require further compromises. -FG: Do we need a new PR? Answer by JCM: Yes - IVOA Working Draft and then PR. -FB: Did you give up on the idea of making STC based on Quantity? Answer by JCM: For now, to get it out, we don't. Answer: Other groups need it now. -FB: Is it impossible to serialise STC in VOTable? Answer: In principle yes, but be very careful. Preference to import the schema. Conclusions: Ready to go to PR.