IVOA Web>IvoaDataModel>ObservationProvenanceDataModel>ProvenanceDataModelLegacy (2016-10-04, KristinRiebe)

Provenance Data Model Legacy

Original concept in Observation data model IVOA note (2005)

The basic concepts of this model have been discussed as early as 2002/2004. It gave birth to the Spectral data model and the characterisation data model. The draft for the full Observation model became an IVOA note in 2005. The concept of provenance was introduced in this note.

Motivation and use cases

The motivation and use cases for a provenance data model have been discussed at most of the Interop meetings between 2006 and 2012 with evolution of the concept:

Interop	Content	Materials
Victoria 2006	Close to completion of char datamodel, motivations to start full Obs data model including Provenance	Spring 2006
Beijing	More detailed proposal for the content	Spring 2007
Cambridge UK	New provenance use cases gathered from a spectroscopy workshop held at ESAC in 2007	Fall 2007
Trieste	Looking for new partners with various interests and observational background.	Spring 2008
Baltimore	Juan de proposes a radio astronomy view of Provenance	Fall 2008 / Juan de Santander Vela
Baltimore	Mireille proposes a core Provenance container with hooking of specific details in addition	Fall 2008 / Mireille Louys
Garching	Juan de moved to ESO and proposes with Arancha a provenance view for ESO requirements	Fall 2009 / Juan de Santander Vela
Garching	Igor is moving to CDS. He proposes provenance use cases for complex datasets	Fall 2009 / Igor Chilingarian

There were splinter meetings on provenance in Baltimore, Garching and Nara. No trace has been left about this. But they have been opportunities to discover requirements from newcomers in the game. The same happened at EURO-VO meetings in the same period.

First xml attempts

Attempts were made to serialize provenance details. The goal was to demonstrate the aibility of the project.

Interop	Content	Materials
Baltimore	First serialization examples in xml. Processing is made of succession of processing stages hooking inpout and output datasets	Fall 2008 François Bonnarel
Strasbourg	An example of xml schema for Observing Configuration has been added	Spring 2009 DM
Strasbourg	Access to Provenance metadata was seen under the "generic data set" concept which gave birth to ObstAP/ObsCOre and DataLink soon after	Spring 2009 DAL
Garching	For the first time Provenance metadata are seen as a complement to ObsCore in the general VO architecture	Fall 2009 François Bonnarel

You can also look directly at the xml schema and the xml example prototyped in this period. They are not consistent however.

UML diagram

According to IVOA rules at that time a Data Model had to be exposed with an UML schema. In 2010 it was considered time to try to build one.

Interop	Content	Materials
Victoria 2010	First (and only) attempt to build a provenance UML diagram	Spring 2010
Nara	In 2010 we were also introducing a use case for VOspace. But we realized progresses were really slow. UML diagram unchanged	Fall 2010

Sao Paulo review (October 2012)

In a dedicated splinter meeting at this interop an attempt was made to relaunch the effort. There was a lot of use cases summarized in the Introduction talk as well as a summary of the previous technological attempts. There were talks by Nikos from Marseille (calculator), ArnoldRots and KristinRiebe, but unfortunately they are lost. A the end of the meeting MarkusDemleitner concluded that the approach followed so far was definitely too ambitious and we had to start with revised ambitions and more stable technology. This started at the Heidelberg meeting in May 2013.

Conclusion

Why was all this so slow and eventually "failed"? In FrancoisBonnarel opinion there was a sociological and a technical reason. The sociological one was that the Provenance task was nobody's priority task (including DM "permanents" like Mireille, Anita, Igor or François) just because there was much more critical VO businness for them at the time. Things have changed now. The technical point on the other side was that there were a lot of classes to develop and many ways were possible to describe the metatada. Probably moving to the W3C core architecture helped to structure the efforts around a sound and stable core.

Attachments

Topic attachments
I	Attachment	History	Action	Size	Date	Who	Comment
xsd	Provenance.xsd	r1	manage	3.0 K	2016-09-16 - 14:06	FrancoisBonnarel	Provenance xml schema
xml	provenance.xml	r1	manage	3.7 K	2016-09-16 - 15:06	FrancoisBonnarel	Provenance xml example document