Hotwired 3 TDIG Breakout

[John Swinbank]

Dear TDIG,

Thanks to everybody who participated in the TDIG breakout session at Hotwired 3 on Wednesday. I think this was a pretty constructive discussion and helps inform where we should be heading in the future. There were about 20 people present, including folks from LSST and other major projects.

The session can be roughly divided into two parts: we first discussed the status of in progress and proposed TDIG standards and notes, then some more long-term “blue skies” discussion. I presented a 5(ish) minute summary of the discussion to the wider conference the following morning: my slides are available at <http://wiki.ivoa.net/internal/IVOA/IvoaVOEvent/2013-11-14_-_Hotwired_3_Breakout.pdf> and I hope they provide a useful overview.

Some highlights:

- We hope to push VOEventRegExt and VOEvent Transport Protocol to the standards process within the next few months. Matthew and I (respectively) will follow up to the mailing list asking for more input on the draft documents shortly.

- Much the same applies to SimpleTimeSeries, except that it will be published as an IVOA note. Again, Matthew will provide an update to the mailing list shortly.

- Mike introduced his proposal for a “VOEventContainer”: essentially a means of bundling multiple VOEvents together with supporting data such as images into a single entity. This could address both issues surrounding bulk transport as well as the stated aim of the LSST folks to include cut-out images with their events. The proposal received a positive response from the audience, but there was some quibbling over technical details. Mike will introduce his proposal and kick off a discussion as to its implementation on the mailing list.

- It was widely agreed that authentication of events will become an increasingly big deal. This is both because subscribers don’t want to find themselves chasing phantoms and because high-profile projects want to be careful of their reputation. We regard pushing this forward as a priority, but there was concern that both of the systems that have seen limited uptake, and that at least part of the reason for that is that XML is a format unsuited to cryptographic signatures.

- At this meeting, as in the past, there was some interest in alternative (to XML) event representations. In particular, we got one direct request for JSON, while there were a few comments that XML wasn’t everybody’s preferred system. Potentially, a non-XML event representation which is intrinsically canonical (ie, there is one and only one way to describe a given event) would alleviate problems with respect to signatures, although note that JSON would not be such a representation. There is an action item against me to investigate potential alternatives and report back: I will do some research here, but would certainly appreciate your input.

Cheers,

John