(1) The search process for the University of Michigan Orphan Works Project was not automated. It was performed simultaneously by two separate researchers whose results had to match. This has been known since the high-level outline of the search workflow was published by the project itself. The steps they performed when attempting to locate the rightsholder may have been inadequate, of course, but they were not automated.

(2) The Authors Guild might have “publicized the list of potential orphan works in HathiTrust’s hopper,” but it is important to note that they were not publishing some leaked internal memo; they were merely rebroadcasting the list published by the project itself, published for the explicit purpose of getting outside experts (such as the AG and the SFWA) and the crowds to review their work and find anything they’d missed. Thank you for doing your part with regards to Fletcher Pratt; I hope that you have some success with his case in the future, because if you continue to be unable to find contact information for the potential rights-holder, his work will continue to be an orphan, and that would be unfortunate.

(3) You’re right that experts such as authors groups should be consulted and their knowledge used in any screening process. Sadly, the Author’s Guild refused to help, instead waiting for the the second stage review list to be published to see if they could catch HathiTrust out with the expert knowledge that they had refused to share.

(4) I know of no evidence in the record suggesting that anyone involved believe that “a work is an orphan simply because several generations have passed since the author died.” They would not have bothered to review any book with a deceased author if they believed something like that, and yet they did perform reviews of such works.

Second, two challenges:

(1) I defy you to prove your comment that the University of Michigan Orphan Works Project “is based on Google’s notion that a thousand random books are worth more than a few carefully chosen ones.” The Orphan Works Project is being led by a university and focusing on university-held collections. I do not think that such academic research collections are in any way random or unimportant.

(2) Allow me to suggest further that students learning their art and academics doing research often need to consult works published between 1923 and 1963 (the initial focus of the project), may need to consult more than one such at a time, are quite numerous, and are spread out across a great many disciplines; it is possible that they will collectively need ready access to “hundreds, if not thousands” of works published across those four decades rather than “a few carefully chosen ones” that would have to somehow satisfy the entirety of academia. If you have a better solution to that need than simply claiming that it’s impossible, I’d like to hear it.