1976 · Free Software Newsletter · Work 967 of 1,871
The cramped back office of a small computer newsletter in Menlo Park in California on a spring evening, on a long wooden trestle table two portable cassette recorders side by side joined by a short black cable, the reels of both turning, a row of freshly copied grey cassettes in clear plastic cases, their shells plain and blank, beside them a stack of thin stapled booklets with plain cream card covers seen edge-on from the side, a glass jar holding a few coins, a hand dropping one cassette into a plain brown envelope, a rolled sleeping bag in the corner, warm lamplight and blue dusk in the window.
What it was used for
Tiny BASIC is a small BASIC interpreter for the first microcomputers, spread through a newsletter that printed source code. The first issue appeared in January 1976 as "Dr. Dobb's Journal of Tiny BASIC Calisthenics & Orthodontia", subtitled "Running Light Without Overbyte", edited by Dennis Allison and Bob Albrecht of the non-profit People's Computer Company in Menlo Park. Its purpose was to give owners of Altair-class machines the source code of a small BASIC interpreter to type in. Tom Pittman gives as the reason for Allison's precursor article "Build Your Own Basic" that the computer kit cost 400 dollars, was hard to program in machine language, and that Altair BASIC at 150 dollars was considered too expensive.
What was new
What stands out is the publishing model: not a finished program for sale but printed source code. Allison specified an intermediate language (IL) to be interpreted by a program in the machine. Tom Pittman writes in retrospect that his was the only version to follow that specification closely, and that the Tiny BASIC machine was completely defined in 443 bytes of IL code against more than 2K for hand-written assembly. He was not first: Dick Whipple and John Arnold came before him. The paper was planned for 3 issues; from issue 2 in February 1976 Jim Warren took over as editor and it became a running magazine. Pittman charged 5 dollars, while the others gave their versions away.
What the sources leave open: Allison's original memory estimate is in the reprint of "Build Your Own BASIC" in issue 1 (Internet Archive): 6 to 8 Kbytes for a minimum BASIC interpreter, 12K for a comfortable system. No figure for it appears in the text. Pittman's statements come from his own retrospective in Dr. Dobb's of January 2006, a participant's recollection 30 years later. The plan for 3 issues is Jeremy Norman's statement.