I will reply again about the dates and computers soon.

For the date 1960 I found:

“The availability of computer technology automated the process of randomizing these permutations. José Férez Kuri’s critical anthology Brion Gysin: Tuning in to the Multimedia Age shows four examples of computer-generated permutation poems, programmed to appear in block formation by Ian Somerville in 1960.”
Funkhouser, Christopher. Modern and Contemporary Poetics : Prehistoric Digital Poetry : An Archaeology of Forms, 1959-1995. Tuscaloosa, AL, USA: University of Alabama Press, 2007:

This is repeated in several sources, including the New Museum catalog, none of them mentioning the computer type (except a comment in the source code for the first re-creation for the New Museum Show itself: https://github.com/josephmoore/Permutations/blob/master/ABOUT.txt

For the Honeywell reference, I found varying dates (one even claimed 1959).
However, 1965 seemed to pop up frequently. I am not sure where this first came up. Something tells me that this is more likely.

Sommerville helped “puis avec d’autres de ses poèmes, cette foi-ci à l’aide d’un ordinateur Honeywell, en 1965.”
Jaques Donguy, “Cyberpoésie” in Zigzag poésie: formes et mouvements, ed. Frank Smith, Christophe Fauchon (Paris: Flammarion 2001), 176.

However, there are especially dumbfounding (or just dumb) claims like this that have the 1965 date:
“In Zusammenarbeit mit dem Mathematiker Ian Sommerville realisiert Brion Gysin das Gedicht
“I am that I am”, das 120mal nach einer mathematischen Formel permutiert wird. 1965 gelingt den beiden Autoren eine Permutation bis ins Unendliche durch den Computer Honeywell.”
Saskia Reither, “Computerpoesie: Studien zur Modifikation poetischer Texte durch den Computer (Bielefeld: Transcript, 2003), 120

At least one thing I can add: It’s always Sommerville, with two “m,” not one.

http://alpha60.de/research/there_must_be_an_angel/DavidLink_MustBeAnAngel_2006.pdf

there’s a picture on page 18 of a schematic for the program, with a date given as June 1952. Given software development at the time, I think it’s reasonable to stick with the 1952 as the year of writing (even if he wasn’t running it regularly until 1953).

I know the online version of David Link’s emulator well. I used to teach using it until it was taken offline. The emulator page made the distinction you’re talking about, but the first text on that page was the line “Welcome to the homepage of the Manchester Mark I emulator by David Link!”

I do understand what you’re saying about distinguishing these computers, but the first article you mentioned in comment 1 uses “University Mark 1” and “Ferranti Mark 1” and doesn’t use “Manchester Mark I” (or “1”) at all.

And an in-depth article that isn’t online, Noah Wardrip-Fruin’s “Manchester Mark “Digital Media Archaeology: Interpreting Computational Processes” in Media Archaeology: Approaches, Applications, and Implications. Erkki Huhtamo and Jussi Parikka, eds., Berkeley: University of California Press, 2011, uses “Manchester Mark I” to designate the computer.

Furthermore the program itself refers to the computer that it’s running on as “M.U.C.” (Manchester University Computer).

So, while I’m inclined to say that Ferranti Mark 1 is the best term – and indeed I’ve just changed the page to use this term – it seems that some different terms have been used by scholars for the computer on which the program ran.

To complicate things further, David Link’s article “There Must Be an Angel: On the Beginnings of the Arithmetics of Rays” cites a personal communication that the love letters were posted after August 1953 and before Turing’s death in May 1954. That seems to suggest that the program might not have been written in its final, documented form until 1953, although perhaps the algorithm is earlier. The date could possibly even be 1954, but that seems unlikely to me given that Stratchey’s article about the system in the journal Encounter appeared in October 1954. It takes a while for a project to be done, to write about it, and to have an article published about it. The article does not mention any dates, though, except for the 1951 census.

Do you think the evidence suggests that I should use “1953” or “1952?” for the date?

Here’s for example a site that runs the Love Letters program
https://web.archive.org/web/20130704144157/http://www.alpha60.de/research/muc/

The emulator currently runs Christopher Strachey’s “Loveletters” program from 1952 in its original form.

The Mark 1’s forerunner was the “baby” machine that executed its first program in June, 1948. Then, the Manchester Mark I prototype was built and used from April 1949 to August 1950. In February, 1951, it was replaced by the industrially manufactured Ferranti Mark I that actually ran the loveletter program.

“Kick That Habit Man” is shown in Back in No Time as written in 1959, broadcast on the BBC in 1960. Other sources, including Chris Funkhouser, say his program was written in 1960.

However, many sources name the Honeywell Series 200 Model 120 as the computer that was used, and this computer was not introduced until 1965. Honeywell’s first computer was only developed in 1958 and installed for the first time in 1960.

So, what I have now can’t be right, but is the date of Somerville’s program later or is it a different computer that was used?

I think that term may have been used to refer to the Ferranti Mark 1.

HB, I do believe I just got the 1960 date off the Gysin reader’s date for “Kick That Habit Man” specifically. I’m aware that there were poems with random order and read some about the “random sequence generator” that was used. I was more interested in the (seemingly) deterministic permutation, and tried for a bit to get my system to work in the same way as with Gysin’s “Kick That Habit Man.” That’s still my particular focus, but an option to shuffle the lines or not might be neat. I will look into the date issue soon, in the next few days.

The University Mark 1 provided a computing service up to the autumn 1950, when it was dismantled and scrapped. The first Ferranti Mark 1 was delivered to the University in February 1951, where it resumed the provision of a computing service.