Eliza OCR Project

This is an offshoot of the Eliza 6000 project.

When I needed a copy of the original Eliza script, every copy that I could find on the Internet had OCR errors or deliberate modifications. After more than 50 years, someone should make an accurate copy!

This project's goal is to produce a completely accurate text file copy of the script, as it was published in the January 1966 CACM. You can help find the last remaining OCR errors (or help verify that there are no remaining OCR errors) by playing with the Eliza OCR Checker web app, which uses what I call “the clip-&-sort trick” to make OCR errors more conspicuous.

This project shares the fora, bug tracking system and source code repositories of the Eliza 6000 project.

In the repositories, use the “CACM-1966” bookmark to get the script as it was printed in 1966, without any of my modifications. You may also be interested in the “MIT-1965” bookmark, which tracks my fixes of transcription errors that occurred when the CACM paper was published.


You probably don't want to download the web app (you can run it just by visiting it) but if you do want to download it, you can use the command line “wget -nd -P eliza-ocr-checker -r -k -np https://eliza-6000.munyer.com/ocr-checker/”.

Acronyms used above: ACM = Association for Computing Machinery; CACM = Communications of the ACM; MIT = Massachusetts Institute of Technology; OCR = Optical Character Recognition.