1973 User Study Notes
Overview
In 1973, at PARC [then Xerox PARC], I organized and helped to run two “studies related to secretarial text editing”. The typewriter era was in full swing. In virtually every business, secretaries and dedicated typists did most of the typing. Managers dictated letters and memos. Then they marked up drafts that the secretary retyped with the indicated changes.
At the time, user studies, when performed at all, were called “user tests”, an unfortunate term that made it sound like the user, not the software, was being tested. To avoid that connotation, we called our studies interviews, not much better because the term failed to capture the observational aspect.
The most frequent participant in our 1973 studies was Sylvia Amundsen [then Adams]. Before joining PARC, Sylvia had prior experience using IBM’s ATS (Administrative Terminal System) to prepare, edit and manage engineering documents. We also ran studies of other users who had less experience with word processing software than Sylvia or no experience at all. The notes described on this site are all from sessions with Sylvia. They were originally all handwritten.
Abbreviations in the original notes:
A = append (a CGEDIT command)
Appd Char = append character (a CGEDIT command)
AP = append paragraph (a CGEDIT command)
BS = backspace (like [backward] delete on today’s keyboards)
CA = command accept (commits a command) (available as both a key and a mouse button)
CD = command delete (backs up a step in command specification) (both key and mouse button)
D = delete (a CGEDIT command)
J = jump [instantly scroll] the indicated line to the top of the screen (a CGEDIT command)
JB = jump to the beginning of the document (a CGEDIT command)
JE = jump to the end of the document (a CGEDIT command)
LF = line feed (used in the notes from Study 2 Day 2 to mean “start a new line of text”)
R = replace (a CGEDIT command)
WD = William (Bill) Duvall
WF = write file (a CGEDIT command)
Also, in CGEDIT:
To “bug” something was to point to it with the mouse cursor and click the first mouse button.
Text was only selected during a command, not between commands; the selections (there could be two) were highlighted by underlining them.
The term “levels” referred to the levels of an outline; to “adjust a level” was to indent/outdent a portion of the outline.
As on typewriters of the day, what we now call “caps lock” was called “shift lock”.
First Study
The first study took place on May 30, 1973. I collaborated with two visiting researchers: technology writer Nilo Lindgren and SRI computer scientist Barbara Grosz [then Deutsch]. Barbara, who later became a C.S. professor at Harvard, had more experience than I running user studies. We decided to run a “blank screen study” in hopes of determining what Sylvia found “natural” or “intuitive”.
Notes from the First Study (as transcribed for legibility on July 9, 1973)*
Second Study
The second study ran from August 8-13, 1973. I collaborated with Bill Duvall, who had created CGEDIT (also known as NCG), a subset of Doug Engelbart’s NLS. His subset was focused on outlining and general text editing. On August 8 and 9, Bill and I gave Sylvia lessons in the use of CGEDIT while I took notes. On August 13, I conducted a follow-up interview with Sylvia.
Notes from the Second Study Day 1*
Notes from the Second Study Day 2*
Notes from the Second Study Interview*
Mode issues surfaced on Day 2 and during the interview, e.g.:
Difference between append and insert forgotten.
“What do I do to put a period in there?”
“It took too many strokes to accomplish something as simple as a delete character.” Instead of d c <point> <bug> <ca>, she would have liked <point> <delete>. “To append a [new] character, simply point at the last [existing] character and type the [new] character.”
“It was confusing that BS (the <delete> key) deleted to the left of the cursor [during text entry] but delete character [d c] deleted the character at the cursor [when the d command executed].”
“I would like to point at [an error I just made] and correct it.”
Synopsis
Some time between August 13 and November 1, 1973, I prepared a synopsis of the two studies that I later drew upon in my final report about the studies.
N.B.: The final page of the synopsis says, “Like insert mode in CGEDIT: don’t have to make space first”. What Sylvia liked about “insert mode” was clearly not “being in a mode”. She liked the automatic rewrap that occurred whenever she added text to a paragraph.
Synopsis of the two 1973 studies*
* The original screen shots posted on this site were lost. This page will be updated if new links are found.