prov-wg: Report on HCLS call

Hi All,

Yolanda and I gave a presentation to the HCLS Sci-Discourse group 
yesterday about the current status of the group and gave a brief 
overview of the core model. Below are some notes Yolanda made. You can 
find the slides I used here : 
http://www.few.vu.nl/~pgroth/prov-overview-update-pg.pdf

I will update these with the appropriate logos etc and place them on the 
homepage.

My take aways:

- They really like the simplicity of the model.
- They are keen getting implementations
- They like the idea of extensibility.


Regards
Paul

-----------------------------------------------------------------------------------

NOTES ON PROV PRESENTATION TO HCLS DISCOURSE GROUP,Feb 20, 2012

PaulGroth presented slides:http://www.few.vu.nl/~pgroth/prov-overview-update-pg.pdf

GullyBurns: I would want to use a direct link between activities, since there may be many intermediate products generated and it is very handy in our work.  PaulGroth pointed out wasInformedBy.

TimClark: The slides are very good and should be used in our intro documentation.

TimClark: We started working with Susanna Sansone on an OWL model to encapsulate ISA-TAB data for lab experiments.  We started saying "we need to subclass OBI" but then we migrated to using PROV.  It is very simple to express experiments using PROV.  The simplicity of the PROV model was very important.  We have been working with DanielGarijo and YolandaGil on this.

EricPrudHommeaux: Asked for more details on PROV-AQ.

DavidShotton on IRC: It would be useful to distinguish the STATUS of a document as an output from the ROLE of an AGENT

DavidShotton: Great work, simplicity is very important.  It would be useful to have time-dependent roles.

PaulGroth: Regarding time-dependent roles, they add complexity to the model.  We've discussed two broad uses for PROV.  One is to annotate simple provenance, for example if I've got a web doc or publication and i want to assign some provenance info.  A second case is tracking provenance in an automated system.  In this second case, there may be fixed or time-dependent things (egPaul on Feb 20th at 16:43  in Amsterdam).  Supporting both of these views has been a challenge to the group.  We think we've got it with this model of starting with a simple model but having additions you can use.

AnitaDeWaard: Are there current implementations?

PaulGroth:  We have some initial implementations.  In Taverna, as part of Workflows4ever, they have represented all their provenance with PROV.  In Wings as well.

TimClark: It would be useful to use provenance to model the process that a paper goes through during submission and publication.

AnitaDeWaard: Publishing the workflow as part of a paper is very important, something that people are looking at including Yolanda.

YolandaGil: The workflow involved in publishing a paper that Tim mentions is an interesting one, different from the workflow that represents the method of the paper that Anita mentions.  But the latter has a role to play in the former.  Representing the provenance of the results of a paper is important to reviewers, who can check the validity of the paper.  Another important aspect of provenance with respect to publications is to be able to assign credit, which is very important to scientists.  Every paper could eventually have something similar to "movie credits", where everyone who participated is acknowledged for their role even if small.

AnitaDeWaard: Regarding credits, it may be useful to relate PROV to ORCID:http://about.orcid.org/

TimClark: There will be a workshop in May in Cambridge.  Actually two workshops.

AnitaDeWaard:  Thank you, we should have another discussion on PROV in a few months, when the WG has progressed more.

Received on Tuesday, 21 February 2012 12:17:47 UTC