As it turns out, the analysis glitch led us to all believe that the source was 300 times brighter than it actually was. As the circular I quoted above said, there is a known X-ray source in that particular globular cluster, but when an object suddenly appears 300 times brighter, something awesome is usually going on (provided the analysis was correct).TITLE: GCN CIRCULAR NUMBER: 16336 SUBJECT: Swift trigger 600114 is not an outbursting X-ray source DATE: 14/05/28 07:57:12 GMT FROM: Kim Page at U.of Leicester <firstname.lastname@example.org> K.L. Page, P.A. Evans (U. Leicester), D.N. Burrows (PSU), V. D'Elia (ASDC) and A. Maselli (INAF-IASFPA) report on behalf of the Swift-XRT team: We have re-analysed the prompt XRT data on Swift trigger 600114 (GCN Circ. 16332), taking advantage of the event data. The initial count rate given in GCN Circ. 16332 was based on raw data fromthe full field of view, without X-ray event detection, and therefore may have been affected by other sources in M31, as well as background hot pixels. Analysis of the event data (not fully available at the time of the initial circular) shows the count rate of the X-ray source identified in GCN Circ. 16332 to have been 0.065 +/- 0.012 count s^-1, consistent with the previous observations of this source [see the 1SXPS catalogue (Evans et al. 2014): http://www.swift.ac.uk/1SXPS/1SXPS%20J004143.1%2B413420]. We therefore do not believe this source to be in outburst. Instead, it was a serendipitous constant source in the field of view of a BAT subthreshold trigger. This circular is an official product of the Swift-XRT team.
So.. what really went wrong?
As you may remember, last night, NASA Goddard Space Flight Center's Internet was knocked out by a storm. Therefore, for a very long time, we only had access to the little data that people had grabbed before it went down, which really wasn't much. This meant that one of THE hallmarks of science was unable to happen last night; no one else was able to look at and re-analyze the data beyond what the Swift pipeline had already done. Without that, we all trusted the Swift data analysis pipeline, in part because it was the only data we had available.
Data analysis pipelines are fully automated computer programs; raw data comes in, analyzed data comes out. Only this time, it seems there was a miscommunication. Computer routines are never perfect, just ask anyone who's ever written one. They can be filled with insidious little bugs that will never show their faces until some very specific moment, so the program will work fine otherwise. So with limited data available, and the data that we did have being flawed, we were led to the wrong conclusions about this source. Phil Evans, an astronomer who works on the Swift team says he was able to identify the error and correct it.
So why didn't we all sit on our hands and wait for the full data to come out? Well, that's an easy one. When something BIG happens to astronomy, the entire field moves very fast (there aren't too many of us, so not much inertia). Basically, we rush to try to get as many different eyes on the source as possible. This involves rapid dissemination of information, which astronomers on social media (like a lot of us are) are very good at doing. A gamma ray burst only 2.5 million light years away would have been a HUGE deal, likely one of the most important astronomical events of the decade. So we moved even faster than usual.
As Dr. Rutledge pointed out, this episode has been a pretty fascinating look into how science works. As soon as someone was able to re-analyze the data, they did so and issued a correction which is now making its way through the field, in most cases with a sign of disappointment, and some tongue-in-cheek humor.