Letter O considered harmful? [closed]

Question

Back in the day, the FORTRAN standards committee reviewed a technical proposal called "Letter O considered harmful". I used to be able to find a link to the text of this proposal on the net, but it seems to have disappeared since the last time I looked for it -- the link disappeared off the relevant Wikipedia page and the only Google hits for the term are references back to Wikipedia. Does anyone happen to know a good repository of information about FORTRAN so that I could track it down, or even better, have a link to the proposal itself?

Answer 1

You are indeed correct!

Yes, there was such a proposal (entitled "Letter 'O' Considered Harmful") in the official set of documents supplied to voting members at the November 1976 meeting of X3J3 that was held at Brookhaven National Laboratory. (At this same meeting, committee chose "Fortran 77", with six lower-case letters, as the name for this revision of the language.)

I am able to verify this because I was not only the host for this meeting but also the actual author of this anonymous "proposal". As such, I enlisted the typist (my boss' secretary, Bette) to type up this phony "proposal" in the proper format and slip it into the official distribution provided at the meetingplace (Conference Room B of Berkner Hall).

Loren Meissner was so amused by it that he wrote a little item in a Fortran publication for which he was editor. Walt Brainerd also mentioned it in his publication. I had sworn both of them to secrecy regarding my little prank, so those articles did not identify me. (Sorry, I don't recall the names of these two publications.)

The lists of pro and con arguments (as was typical of X3J3 proposals in those days) included:

• Restoring the number of Fortran characters to 48 (by omitting 'O'to counterbalance the addition of the colon ':')
• Solving ambiguities caused by nested DO loops.
• Eliminating problems with (deprecated) Hollerith specifications in FORMAT statements.
• Preventing misuse of GO TO statements.

while the "con" list contained only one objection (with a disclaimer):

• This proposal may invalidate some existing FORTRAN programs, but most of these are probably "standard-conforming" anyway.