Correctly identify which pipeline has made low-latency detections vs medium-latency follow-up
In O3a, the pipelines credited with observing candidates in low latency included pipelines that did not automatically observe the candidate but instead saw something consistent in after-the-fact tests performed during the RRT call before sending the circular. When I've discussed this with astronomers, they were shocked, and felt (in my opinion rightfully so) that the text in the circular was misleading. I believe it is important we make a distinction between the two situations. The only pipelines that should be credited as observing low latency candidates are the pipelines that had automated gracedb uploads at the time of the candidate. We can, and should, mention that offline follow-up showed consistent results in other pipelines, but equating the two processes is misleading.