InterOp Roundup

[John Swinbank]

Dear TDIG,

Thanks to those who made it to our TDIG session at last week’s InterOp: the biggest TDIG session yet. Thanks also — and especially — to Sébastien and Petr for their excellent talks. After such a successful meeting, the challenge now is to maintain that momentum and make sure we get things done.

First order of business is to thank Mike Fitzpatrick for his excellent work over the last three years as vice chair of the group, and to welcome Dave Morris as his replacement. My own position as chair has been extended for one further year: I will step down at the Spring 2017 InterOp.

Now, on to more technical matters.

# VOEvent

As you know, the VOEvent Transport Protocol is currently a proposed recommendation — http://www.ivoa.net/documents/VOEventTransport/20160503/index.html — and will be entering an RFC period shortly. I gave a short presentation in Stellenbosch summarizing the purpose and design of the protocol, the current state of the document and available implementations. Slides are here: http://wiki.ivoa.net/internal/IVOA/InterOpMay2016-TDIG/2016-05_-_VTP_at_InterOp.pdf. If you have questions, concerns or contributions before the RFC opens, please do feel free to follow up to this list.

One of the outcomes from the focus sessions was that new large facilities — in particular LSST — are concerned about how well existing VOEvent infrastructure scales to their requirements. David Ciardi of LSST announced an LSST Corporation funded project to study how we might go about evolving current systems to meet their needs (http://wiki.ivoa.net/internal/IVOA/InterOpMay2016Focus/LSST_IVOA_20160506c.pdf pages 11–13). Exactly what form that study will take remains to be seen, but it will be important for us as an IG to engage with that process, and also — if necessary — to ensure that we represent the needs of the wider community. To that end, I invite those of you with thoughts on the future of VOEvent to mull over the issues and start pitching in with your thoughts on the mailing list. Maybe useful to read some thoughts that Tim Staley and I wrote at the time of last year’s Hotwired meeting to get the ball rolling: http://voevent.readthedocs.io/en/latest/future.html.

# Complex Time Domain Data

Discussion has been swirling on this mailing list for some time about how to represent time series, and it certainly seems that this is a big deal to the focus session projects: it was one of the top issues that arose in the focus sessions (see the summary at http://wiki.ivoa.net/internal/IVOA/InterOpMay2016Focus/FocusSessionSummary.ppt). In our TDIG meeting, Sébastien talked about how heterogenous even “simple” time series data available in Vizier is, clearly demonstrating some of the challenges we faced. Petr demonstrated a potential solution based on the cube DM (http://www.ivoa.net/documents/NDimCubeDM/index.html). While Petr’s approach isn’t yet polished, and hasn’t yet been demonstrated to solve all the issues, it’s an impressive step forwards and clearly points to a potential way of making progress.

Where do we go from here? I think our first priority is to ensure that we have a comprehensive set of use cases from which we can derive requirements. Enrique has made a good start at this here: http://wiki.ivoa.net/twiki/bin/view/IVOA/CSPTimeSeries. I invite TDIG members to provide additional use cases, or augment those that have already been defined. I’m also calling for a volunteer to act as an editor to bash everybody’s use cases into shape and ensure we present them in a nice, coherent way (as Enrique’s already are). Any takers? I’ll also reach out to the focus session projects, and directly solicit use cases from them that we can incorporate into our thinking.

Once we feel that’s converging onto some real requirements, we can open a discussion with the Data Model WG to see how we can push this forward to create a really usable standard.

Thoughts? Suggestions? Comments? What have I missed?

Cheers,

John