Copyright © 2012-2013,2014 by Thomas E. Dickey

Comments about Scribe

I encountered Scribe more than once:

Leive was commenting because while both his progress and mine were slow, we had at least a topic. Brian had none, and his advisor had a job in store to help Brian put something together. As it turned out, it was a better "pub" program.

I first noticed Scribe as its user manuals appeared in the bookstore. There was more than one edition. I have at hand these:

Scribe Introductory User's Manual, 2nd Ed., July 25, 1979
This is copyright by Brian Reid, but lists (as does the third edition) Janet H. Walker as co-author.
Scribe Introductory User's Manual, 3rd Ed., May 1980
The third edition gives August 1978 as the date for the first edition. This version is copyright 1980 by Unilogic Ltd.

The first edition (which I discarded long ago, having need for only two Scribe manuals) gave a different story. It said that the program's development had been sponsored (paid for) by Rome Air Development Center (RADC). The later editions do not mention this, and it seems that there is nowhere a useful source explaining how it became transformed from government sponsored research into private property. For those interested, the first edition is available via NTIS which says the contract number was

F44620-73-C-0074 DAAG29-77-C-0034

Around the time that the third edition was published, some of my coworkers were setting up a VAX/VMS system. It was paid for by our division, though ultimately my department had no benefit from it. The "owners" of the VAX system were largely people whom I had dealt with as a graduate student, starting with the printer diverted for a Unix system.

In my department, there were people using a remote timesharing system to run Basic programs. I thought that if we could move those people onto the VAX, we would get some return on the money spent. The people running the VAX had different ideas; they were uninterested. Instead they planned to use the machine for word processing. At the time, the division used a pool of secretaries with NBI workstations, but most typists used IBM Selectrics. Engineers, for the most part, did not type.

They intended using Scribe. I was told the license cost was $20,000 (very high), but on suggesting that they use (or at least evaluate) the free version was told that was unacceptable, that it was "full of bugs". Further discussion led to the assertion that the free version was not available.

Later, in 1984 I wrote a program "describe" (prompted by "deroff" for "roff") which stripped markup information from a Scribe document. At that point in time, Scribe required license keys to run (as I recall, 20 alphanumeric characters), but it was someone else's problem by then.

In writing this, I am reminded that Unilogic was founded by Michael Shamos (and others), Shamos vetted a weakly designed experiment used by one of my peers in his research (28 independent variables with 43 data points).

My peer on the other hand chose to attack my own experiment (which was not as weak), and in particular based his criticism (given the numbers that he cited) on the Sackman report.

In researching to determine the actual facts in this (and ultimately that his criticism had no merit), I came across a comment in one of the trade magazines. It quoted someone as saying that at MIT, writing a compiler is what masters students do. Proving that a given compiler is good in some respect (analysis) is what PhD students do.

Scribe was a nice program ("was", because no one appears to use it anymore), but there was no analysis involved.