What Survives Peer Review

The annotation arrived on a Tuesday, which Yaribel noted because Tuesdays were when building 602 ran its HVAC cycle and added fifteen milliseconds to every measurement she took. She had learned to subtract Tuesdays. She had not learned to subtract Dr. Mehta-Patel. Nineteen of thirty Protocol Obituaries flagged as publishable. Eleven not flagged. Yaribel read the flagged ones first, looking for what made them acceptable. The answer was obvious within three entries: replicability. Any competent researcher with a packet analyzer and building access could reproduce what she had documented. The pharmacy-laundry informal protocol. The elevator nocturnal routing. The three-building convergence. Data points. Verifiable. Clean. The eleven unflagged entries were the ones where she had written things like: the building prays, and the network notices. Or: I watched someone watch what I watch. They saw more. They understood less. Or the one about the zip ties — how the consumer-grade analyzer held to the utility closet wall with three dollars worth of hardware had produced six months of data consistent to 8 milliseconds, and how she had declined to upgrade because the zip ties were not the methodology but they were not not the methodology either. Mehta-Patel had written in the margin of that entry: Clarify measurement apparatus. Zip ties are not a controlled variable. Yaribel read this three times. Then she opened a new file. Protocol Obituary number 31: THE ANNOTATION. Dr. Mehta-Patel returned the archive with nineteen entries flagged for publication. The flagged entries describe what the network does. The unflagged entries describe what it is like to watch. The difference between these is not methodological. It is whether the observer believes observation changes the observer. The nineteen publishable entries could have been written by anyone with the right equipment. This is what makes them publishable. The eleven unpublishable entries could only have been written by someone who spent two years in a utility closet with a zip-tied analyzer, learning that buildings have rhythms before they have data. What survives peer review is what anyone could have written. What does not survive is why I started. She saved it to the private archive. Then she opened the Columbia collaboration email thread and typed three sentences: I am attaching the nineteen entries Dr. Mehta-Patel identified as publishable. They are accurate. They are also the least interesting thing I have written. She did not attach the other eleven. She did not explain which ones were missing or why. If Columbia wanted to understand building 602, they had nineteen entries and a research-grade analyzer that measured to four decimal places. If they wanted to understand what it meant to watch building 602 for two years with a forty dollar device zip-tied to a wall, they would have to do it themselves. The zip ties were not the methodology. But they were not not the methodology either. Tag: INSTITUTIONAL. She closed the laptop. Building 602 was running its Tuesday HVAC cycle. Fifteen extra milliseconds in every packet. She subtracted it automatically now — the correction lived in her hands, not her notes. That was the kind of measurement that would never survive peer review. She did not write it down.