*IVOA DAL WG Running Meeting #3 Wednesday 2 September 2020 - 13:00 UTC - vconf Participants: (13) Marco Molinaro, James Dempsey, Laurent Michel, Markus Demleitner, Renaud Savalle, François Bonnarel, Gregory Mantelet, Ada Nebot, Jon Juaristi Campillo, Mark Taylor, Patrick Dowler, Tom Donaldson, Dave Morris *Agenda: 1. - DataLink templated endpoints (continued) 1. - ConeSearch document progress 1. - UDF EN discussion 1. - TAP & DALI status (added while running) * *Notes *1 - DataLink Examples of RFC 6570 templates URL : https://github.com/ivoa-std/DataLink/issues/27 Examples show RFC 6570 levels 1 &3. [LM] level 4 not useful for DataLink [LM] RFC 6570 only allows use of PARAM value, propose extending the syntax to add a "Fields" syntax to reference data in the VOTable result [FB] Either the value will come from the user or form a field. [MT] Why do we need the extension? The RFC doesn't say where the value comes from. [LM] Need to be able to say where to look [MT] Could just use a field or a param with the listed name rather than using the Fields syntax. Would need a priority in the case of both a field and a pram having the same name. [MD] Also agree that we shouldn't need the Fields syntax [MD] When do you escape values? [LM] This is defined well in the RFC [MD] Are there missing +s in front of the parameters in the examples? [LM] yes [MD] Would like to see a real example, can't see that having these in a separate resource is preferable to having them in the links resource. It's always better if there's just one way to do a given thing. We should have a really strong case if we introduce a second way. [FB] the templating is more flexible and allows interaction with the user through interface. In the links response would have to be replaced by a potential long list of fixed links. [PD] Level 1 examples - difference from DataLink 1.0 is the curly brace structure but can define parameters to be appended at the end. As a result there is an incompatibility between versions and so clients would need to be able to determine which version of DataLink the server is providing to determine how to treat PARAMs and value substitution. [LM] RFC is needed if you want to put items within the pathname. Agree that clients will need to detect the presence of template syntax in the URL, or have a switch for templating mode. FB: Beside template do not forget 3 changes for 1.1 - modify point related to OVERFLOW - add a new FIELD for typing content of the link (dataproduct_type or object type or whatever) - add a FIELD to give the access status of the link (restricetd or not) [PD] volunteered to write this text [FB] Would like to have a working draft prepared before interop *2 - ConeSearch ( https://github.com/ivoa-std/ConeSearch/issues ) a) 12 open issues (+5 for a major release) UCD1+ change #20 [MM] Would like to discuss how clients would react to this in a minor revision. [FB] Thought the decision was not change the UCDs in a minor version [MT] TOPCAT etc will cope fine with the change in UCDs - will just need to update validator [TD] Most clients he has seen already cope with this as some servers make the mistake of using UCD1.1 mistakenly. Any code that uses VOTable has to undestand UCD 1+ anyway. [MM] Could you add these as comments on the github issue please CATALOG & EPOCH new filter params #39 & #40 [FB] EPOCH is related to coordinate system? [AN] Yes this is for coordinates and not related to the time epoch of the observation [MM] Does the EPOCH imply automatic conversion of coordinates between epochs by the server? [AN] Conversion is likely to be complex, so suggest it be a later feature Discoverability & clienthandshaking #33 #32 [FB] Do we plan to include a DALI style overflow response for large queries [MM] Yes that should ahve been included, please raise an issue * b) Styling for "deprecated" parts of the specification (not covered by issues) [MM] Most are currently in sibling sub-section to the recommended approach [MM] Community feedback and involvement is encouraged *3 - UDF EN ( http://www.ivoa.net/documents/udf-catalogue/20200806/PEN-udf-catalogue-1.0-20200806.html ) [MM] Concerns about the PEN introducing new functions (e.g. MOC functions) [MM] Concern is that is only one implementation of healpix functions [MD] Implementation in both Gaia and DaHCS based on an agreed UDF syntax [MM] Would be good to have this called out in PEN [MD] Would prefer this be noted in the RFC page [MM] Agreed [MD] MOC UDF query was published in an ADASS paper [DM] Would be good to document the testing of the interoperability of the implementations *4 - DALI and TAP status - TAP ported from volute to github [MM] WIll alert DAL list on this development - WD-DALI-1.2 now includes agreed xtype additions - unblock ADQL? [MT] Have concerns about the use of 1 or 2 NaN(s) [PD] Implementing this now and will raise an issue once this is complete for discussion What do you want to do ?New mailCopy What do you want to do ?New mailCopy