This archive exists because the literature on how machine translation developed is scattered across conference proceedings, encyclopaedia entries and newsletters that are no longer distributed. Much of it was published free by its authors and then went out of reach when the hosts that carried it closed.
What the collection holds
Three things. The narrative pages cover the moments that decided the field: the 1949 memorandum, the 1954 demonstration and the 1966 report, set out in a timeline. The publications bibliography reproduces the survey literature. The compendium catalogues the translation systems themselves.
Where the material comes from
The secondary literature is largely the work of W. John Hutchins (1939–2021), a librarian at the University of East Anglia who wrote the standard histories of the field and compiled the catalogue of commercial systems. His papers are reproduced and cited here as sources; where a claim comes from his work, the work is named.
Primary sources — press coverage of the 1954 demonstration, the text of the ALPAC report, contemporary product documentation — are cited to their own publishers wherever they remain available.
Why the addresses do not change
Encyclopaedia entries, conference tutorials and reading lists have been citing these documents for two decades. A reference collection that reorganises itself breaks every one of those citations. Where a paper was published at a particular address, it stays there, and new material is added around it rather than on top of it.
Contact
Corrections, missing papers and citation queries: [email protected].
Is this site affiliated with any institution?
No. It is an independent reference collection.
Can the material be cited?
Yes. Cite the original paper and its original publication venue, both of which are given with every entry in the bibliography.
How is the compendium kept current?
The catalogue is historical rather than current. It records what commercial translation systems were, as documented at the time they were sold, which is the part of the record most at risk of disappearing.