Summary

  • In RFC 1003, Alan Katz argued for an equation interchange representation as more scientists approached electronic mail through NSF networking and the EXPRESS submission project. His March 1987 memo posed questions and a tentative design, rather than issuing a finished standard.
  • He favored a text form resembling EQN: common structures such as subscripts, fractions and matrices would carry their own layout instructions, with coordinate positioning available for exceptions. The target was the displayed equation, not a machine's understanding of mathematics.

Imagine a physicist sending a paper to a colleague. The prose arrives. So does a line of plain characters that ought to become a stacked fraction, a growing parenthesis and a subscript that sits precisely where the eye expects it. At the other end is a typesetter who knows exactly what the page should look like and has no reason to interpret the physics. Suddenly the Internet's simple promise—send a message—has a footnote large enough to swallow the formula.

Alan Katz wrote RFC 1003 in March 1987 as the National Science Foundation expected more scientists to join the network around its supercomputer programme. He pointed to NSF's Experimental Research in Electronic Submission, or EXPRESS, and its plan to move manuscripts and proposals by electronic mail. Scientists had been using ARPANET mail for years. The missing trick was making their mathematics portable. Katz reported that there was no agreed representation standard, and even the thing to be represented was disputed.

One faction already had a practical answer: send the source for TeX, LaTeX or EQN. It was text, so existing mail could carry it, and a text editor could replace every alpha with beta in a sweep. The price was paid at the keyboard. Katz compared writing such an equation to writing a small programme; syntax mistakes were easy, inspection often required printing or a workstation preview, and many professional typesetters knew the desired appearance without knowing the mathematical meaning. The network could move the file. It could not make every person who touched it enjoy programming.

Full-display systems offered the opposite bargain. Xerox STAR and VIEWPOINT showed the equation while it was entered; Hockney's Egg on an IBM PC let characters go almost anywhere, even into chemical diagrams. Interleaf gave similar visual freedom by cut and paste. Yet the representation could be hidden, hardware-specific or proprietary. Where every symbol had to be placed by hand, flexibility turned into clerical labor. The screen was a comfortable place to compose and a poor treaty between incompatible machines.

Symbolic systems such as Macsyma and Reduce knew something about mathematics, but that was more understanding than this job required and limited input to mathematics the system knew.

Katz set out the choices with unusual clarity. A bitmap preserved a picture but consumed storage and resisted editing. A list of characters, fonts and exact positions could draw anything but made routine fractions tedious. A structural notation could say “subscript,” “superscript,” “fraction,” “matrix” or “growing parenthesis” and let software handle the geometry. His tentative favorite was something like EQN plus absolute and relative positioning for the stubborn cases. LaTeX could also serve, he said; he simply found EQN easier to read.

The escape hatch mattered because scientific notation always eventually produces a symbol that declines to sit in the allotted chair.

The design also drew a line between appearance and meaning. A mathematical engine might understand an expression; an interchange format for displayed equations had to let a secretary or typesetter enter what the physicist wanted to show. Katz warned that even a font change could alter mathematical meaning. A good representation therefore had to preserve enough visual structure without demanding that the person at the keyboard solve the theorem.

RFC 1003 ended as a call to define a standard, not the standard itself. Its durable lesson is a division of labor hidden inside the word “electronic.” Mail moved characters; editors created appearance; scientists supplied meaning; typesetters made the page legible. Katz wanted a small common contract between them. Without it, a scientific network could deliver the envelope and leave the equation on the sender's desk.