What is a VOEvent and why would I want one?¶
The transient deluge¶
Systems for detecting astrophysical transients are becoming increasingly ubiquitous and high volume: from space-based gamma-ray monitors like NASA’s Fermi and Swift missions, through present-day ground based surveys in both the optical (e.g. the Catalina Real-Time Transients Survey and the Palomar Transients Factory) and radio (the LOFAR Radio Sky Monitor and ASKAP‘s Variables and Slow Transients survey), to future facilities such as the Square Kilometre Array (SKA) and Large Synoptic Survey Telescope (LSST). LSST alone promises to detect and announce up to 40 million transients per night when its survey begins early in the 2020s.
Obtaining the best scientific results from these facilities requires timely and accurate follow-up observations of the most relevant detections. However, it is obviously humanly impossible—even given a large cohort of graduate students—to read and understand millions of transient alerts per night, let alone to select those which merit further observation. Even were it possible, humans are slow: slow at reading, but, even more so, slow at negotiating with their peers to organize telescope time.
The VOEvent standard¶
It is this problem which VOEvent seeks to alleviate. It provides a standardized, machine-readable means of describing transient celestial events. It enables the author to specify:
- Who they are;
- What they have observed;
- Where and when the observed it;
- How the observations were made;
- Why they think it is of general interest to the community.
In addition, it is possible for them to provide citations to previously announced events, and to refer to other content outside the context of the VOEvent itself. In short, the aim is to provide the reader not just with a notification that something was observed, but to encapsulate as much information about it as is practical, so that they can make an informed decision about whether to perform a follow-up observation.
Let’s emphasize that last sentence again: the recipient of a VOEvent makes a decision about whether to perform any action in response to a VOEvent. The event itself does not constitute a trigger or an instruction to the recipient, and it does not suggest what sort of follow-up might be appropriate. It describes only the event as seen by the observer.
VOEvents are XML documents. XML is a plain text “markup” language: it makes it possible to annotate a document to show its structure in a way that a computer can understand, while also being human readable (for a broad definition of “readable” and a fairly tolerant human). The VOEvent standard defines an XML “schema”, which describes the subset of all possible constructs which it is legal to use in a VOEvent document. As we’ll see later, it’s perfectly possible to read and write VOEvents using your favourite text editor. However, that’s a lot of effort; there is a range of tools and libraries you can deploy to make your life easier. We’ll cover them too.
The VOEvent standard was defined by Seaman et al. (2011) and it has an official stamp of approval from the IVOA. It is currently curated by the IVOA’s Time Domain Interest Group, or TDIG.
It’s important to realise that the VOEvent standard says nothing of any significance about what you should do with a VOEvent once you’ve got one. In particular, it might seem reasonable to assume that once the discoverer of a transient has encapsulated it in a VOEvent document, they might wish to distribute it to the community, but the mechanisms by which they can do that are not defined by the standard. It’s perfectly to do this by any convenient means: e-mail, post, semaphore, ...
This is both a blessing and a curse. It’s good architecture to separate the representation of events from their transmission, and it means the standard can be relatively concise and manageable. Plus there’s a great deal of flexibility here: if a project adopts VOEvent, it doesn’t tie them into some particular networking infrastructure. However, for somebody who just wants to find news on the latest transients, or who just wants to ensure that their events are available to the community, there’s no obvious place to start.
This situation is partially mitigated by the VOEvent Transport Protocol (VTP; Allan & Denny, 2009). This is an IVOA Note, rather than a Recommendation; in other words, it is a suggestion which has not been through the IVOA’s rigorous review process. There is a revised version of this protocol in preparation, with the hope of developing it into a fully-fledged IVOA standard in future. The changes are intentionally being kept minimal, and focus more on resolving ambiguities in the specification than on changing the on-the-wire protocol.
VTP is minimal by design: it builds upon low-level networking primitives to create something that gets the job done. Arguably, a more modern system would be layered on top of higher level standards, like HTTP, which would provide a more robust basis and make it easier to take advantage of existing infrastructure. However, VTP is the nearest we have to a standard at the moment, and it certainly provides mechanisms for sending and receiving streams of VOEvents with relatively low latency. This document will describe how to connect up to and participate in the existing VTP network.
Higher level functionality¶
We conclude this section with a warning. All of the tools described here are fairly low-level: they provide mechanisms for reading, writing and exchanging individual VOEvent packets. This is necessary but not sufficient to service future science projects. We need to build upon these lower level tools to construct event aggregators and filters and classifiers and distribution systems which make it possible for the astronomer to reason about and respond to VOEvents in bulk. While some important work has been done on this (e.g. Williams et al, 2009, Poci et al, 2015), there’s still a long way to go. Maybe after reading this you will be inspired to contribute.
We’ll return to this point later in considering some of the challenges facing VOEvent in the future.