[License-discuss] Early uses of the term open source (was: "Fairness" vs. mission objectives)

[Rick Moen]

I belatedly noticed this subthread from a few days back, which happily
helps angle back to software licensing and OSI:

Quoting Eric S. Raymond (esr at thyrsus.com):

> No, it wasn't.  Believe me, I did a *very* through audit on existing
> usage at the time I proposed the term for general use in 1998.  I
> might be able to do a better one today, but only because search engines
> today have more reach than Alta Vista did.
> 
> "Open source" had, at that time, a dominant meaning derived from spook
> jargon.  It referred to primary intelligence sources that are publicly
> available as opposed to those that must be gathered by covert means.
> 
> There was occasional, very rare usage of "open source" to describe
> software packages. I remember finding two uses of this kind from
> USENET in 1992-1993.  After the fact one of them was pointed out to me
> by a random Internet denizen who as trying to be a friendly critic and
> didn't know I was already aware of the precedent.

I researched this matter in 2004, and posted my results to this mailing
list,  when OSI was challenged as to its right to be custodian of the
term 'open source' in software context by one Mr. Chuck Swiger.  My
findings (cut off at the end of 1990 solely because of Mr. Swiger's
claim of conflicting established usage before that -- but IIRC I got
similar results right up through OSI's founding in 1998):


Quoting Chuck Swiger (chuck at codefab.com):
[...]
> Do your own searches, then.

OK, good idea.  DejaNews (Google Groups) returns 16 results on the
following search:

  Exact phrase "open source" plus the word "software" over the period
  January 1, 1981 (earliest date searchable) through December 31, 1990
  (your choice of year).  Going backwards in date order:



1, 1990-12-10:

    BSD's open source policy meant that user developed software could
    be ported among platforms, which meant their customers saw a much more
    cost effective, leading edge capability combined hardware and software
    platform.

So very, very close.  Maybe even a hit.  The intended reference seems to
be to visibility of CSRG's source code, not specifically to the right to
use it for any purpose and further develop it independently.

By the way, that's Thad P. Floryan on alt.religion.computers, arguing
with Daniel J. Bernstein.  Amusing.


2, 1990-07-25:

   error("cannot open source file for input");


3.  1990-07-12:

   } else /* Cannot open source file. Should not happen. */
       error = "Huh?? Cannot open source??";


4.  1990-05-12:

   Actually, NSA makes a recommendation for each commodity jurisdiction
   determination which is required under category 13B of ITAR.  It took me
   "only" 16 months to obtain the following paragraph (under the
   Freedom of Information Act) to learn that NSA's position (considerably
   more recent than a decade ago) as of February 1987 was:

      "Although software was developed from open source material,
      the application of that information into the subject software
      program contains cryptographic capabilities that are controlled
      under category 13B."

   The Commerce Department took the completely opposite position:

      "There is no military application identified.  The software is
      also written without a military application in mind."

   I therefore agree that "the US Government is not a single monolithic
   organism with completely coordinated, coherent policies."  My primary
   concern is that those policies must comply with the U.S. Constitution
   and thereby allow the free dissemination of open-source/published
   material -- including software (ESPECIALLY FREE SOFTWARE) which is
   developed directly from published algorithms.

This (Tony S. Patti on sci.crypt) _seems_ to be the aforementioned,
long-established _espionage_ sense of the term (open sources).  Very
close, though, as it does address specifically software.


5.  1990-03-01:

     fputs ("can't open source file ",stderr);


6.  1989-11-19:

    error("Can't open source file %s",srcfilename);


7.  1989-10-03:

     I am struck by the lack of any reference to Virus-L, RISKS Forum
     and other INTERNET services which have for years provided we users
     the best available, open source information on the subject of computer
     viruses.


8.   1989-08-01:

    write sys$error "Can't open ""''source'"""


9.   1989-07-25:

    XtError("Cannot open source file in XtDiskSourceCreate");
    [...]


10.   1989-06-30:

   if ((from = open(source, 0)) < 0) {


11.  1989-01-21:

   error("Unable to open source file.");


12.  1988-12-01:

   } else /* Cannot open source file. Should not happen. */
            error = "Huh?? Cannot open source??";


13.  1988-11-27:

    as_perror ("Can't open source file for input", file_name);


14.  1988-01-09:

    write sys$error "Can't open ""''source'"""


15.  1985-12-15:

    Z    if (-1 == (s = open(source,O_RDONLY)))


16.  1984-12-17:

    * Open source file as standard input



Well, that was almost fun.  It allowed me to pretend for a few minutes
that I'm a typical computerist unable to see the larger point because of
obsession with distracting details.  ;->

You see, if we spot Thad Floryan and Tony Patti pride of first place
with their one-time mentions of the term in 1990, all that _really_ means
in the end is that I'd gladly buy each of them a beer in gratitude for
their helping launch the concept, even though they _went nowhere with it_
-- and _nobody did until OSI_.  Bringing us back to the point:

That OSI established the term (in the software sense) remains
self-evident.

> I also see the existence of shades of grey in terms of "open source",
> such as Sun making Java "mostly open but require a compatibility
> suite", or licenses like NetHack and Moria/Angband, which have (or at
> one time had) a "no commercial use/resale" term.

You can certainly _call_ those open source.  And then I'll politely ask
you to please correct said error if you go around, say, putting them on
the front page of a Web site for "open source developers".  For reasons
already mentioned.  [RM comment:  Mr. Swiger was one of a succession of
critics who had -- then in the early 2000s -- visited this mailing list,
declared OSI to have no authority, and proposed deploying Web sites
redefining the term 'open source' differently.]