[ih] XEROX/PUP and Commercialization (was Re: FYI - Gordon Crovitz/WSJ on "Who Really Invented the Internet?")

[Dave Crocker]

On 7/30/2012 6:52 AM, Vint Cerf wrote:
> the unexpected about-face towards OSI by DOD raised my blood pressure.
> I am reasonably sure this came from the DCA people who touted X.25 and
> lambasted datagrams and TCP/IP. I am not sure why NIST too the OSI
> path but I guess the idea that it was an ISO standards effort led some
> to think it was unlikely that a DOD-developed protocol would ever be
> acceptable to the rest of the world while the cold war was one, etc.


By most measures, the OSI work had more than enough formal thrust to 
fly.  Governments and major corporations were all thoroughly committed 
to it, for a very long time and with quite a lot of money.[1]

We ought to learn lessons from its failure.

While some of us easily rattle of our favorite lists of reasons we think 
the Internet stuff won[2], I don't recall seeing a careful analysis by a 
professional historian, reporter, sociologist, economist or the like. 
The importance of such an analysis should be to provide a cautionary 
tale to those who think that powerful, formal support ensures success.

d/


[1]  In the late 80s, I managed a shop that did TCP/IP stacks, including 
for DEC's VMS.  25% of my revenue came from customers in Europe, the 
supposed hot-bed for OSI.  What I concluded was the OSI had established 
the market need, but hadn't satisfied it. In fact, one of my customers 
ran the IT operation for ISO!  I asked him about the irony of this and 
with no smile he said he was assigned a job to do and his only concern 
was making sure it got done.  He could do it with TCP/IP and he couldn't 
do it with OSI.  I think it was his statement that convinced me which 
side would win this particular war.

[2]  Mine attributes the primary reason to the Internet's having 
operational end-to-end service that was useful.  Actual practice beats 
theories about future use and utility any day.  I think that the recent 
comment on the list about the flat-fee pricing structure is also likely 
correct.

-- 
  Dave Crocker
  Brandenburg InternetWorking
  bbiw.net