[ih] Peter Salus / Baran's work

[John Day]

> On Jan 13, 2015, at 09:48, Noel Chiappa <jnc at mercury.lcs.mit.edu> wrote:
> 
>> From: Brian E Carpenter
> 
>> So, the meme is indeed puzzling and far from universal.
> 
> I've been puzzling over this a bit, and I wonder if the problem isn't a
> conflation between 'we didn't know about it' and 'it must not have been
> talked about'. It's certainly clear that in the beginning the US people were
> not aware of Baran's work - but I wonder if they just assumed that they
> hadn't heard of it because it wasn't talked about?

It is true that for most of us asked to join the ARPANET, the focus was the *ARPANET.*  It is also true that most were computer scientists and read ACM, etc. (in those days). IEEE Trans on Communications was very much concerned with physical layer issues.  (Remember until the late 70s, a data comm textbook was 300 pages of physical layer and 50 pages on everything else. That isn’t a big exaggeration.)

The origin of the basic idea wasn’t a big concern.  Everyone was focused more on what to do with it. (Remember, very few people were really dealing with “network issues.”  Kleinrock’s group was building it, BBN was building it, and Roberts was overseeing it. For everyone else, it was what to do with 1822 and what should NCP look like. As I have said, the fact that it was “packet switching” was not that big a deal to us, it looked pretty straightforward approach to solving the problem. But we were computer people not telephony people.  For us, it was really more a distributed computing problem.  If you are computer person, the idea of picking up a buffer, throwing a header on it and sending it is pretty straightforward.  Admittedly it is a big deal to telephony types.  “Packet switching” per se was invented independently (and in somewhat different forms) by more than Davies and Baran.  It turns out that stat muxes were packet switches at a finer granularity. I think Datapoint also independently came up with the idea, and probably others as well.

> 
> I wonder if that was because computer people (who were the group who
> eventually took up Baran's ideas) may not have read a 'Communication' journal
> (which was possibly focused towards telephone/etc people)?
> 
> And when they did find out about it, the form they got it was in the RAND
> reports, which they might have assumed were not widely distributed, or
> something - therefore not even realizing, at that stage, that it had been
> previously described in the open professional literature.

I would guess that their first exposure was in the BBN reports.  (Remember this was a very small group of people and very focused (as they still are) on what they had to build.) You didn’t have to read Baran’s article or the RAND reports to get the idea. Someone could just tell you about it and you (being a fairly bright person) could fill in the details and look at the BBN reports to see what you needed to do.  The RAND reports were not all that available. (Remember this was the days before documents were online and xeroxing was still expensive.)
> 
> It is certainly ironic that the US people (Taylor, Roberts, etc) had to learn
> about Baran's work by way of Davies and Scantlebury, from another country -
> via someone at the British MoD - it's too bad we don't know who that was.
> 
Now at this level it may well have been a different story and I believe the evidence indicates that those guys did know about Baran.

>> (BTW, there is some discussion of the military vs civilian origins of
>> the ARPANET project in Walter Isaacson's recent book "The Innovators."
>> Not to mention an interesting discussion of Kleinrock's contribution vs
>> Baran and Davies.)
> 
I think the nuclear war meme is really more tightly associated with the Internet than the ARPANET.  Remember when the ARPANET started, it was the height of the Vietnam War and the DoD was not very popular on college campuses.  In fact, I have heard tell there was initial reticence to being on the ARPANET because it was DoD, which lead to John Melvin’s famous line about ARPA, “It’s okay, our money is only bloody on one side.”  ;-)

In 1970, it was far too early to talk about military use, since they didn’t know if it would work. Although it was pretty obvious that tif it did, it would have military uses. One would have not have gotten very far with the argument to build this purely for military use.  And also remember this was before the Muskie Amendment and ARPA was funding fundamental research, like NLS, the ARPANET, IlliacIV, etc.  

I think there is too much tendency in this discussion to think that things back then are like they are now. They weren’t.

Take care,
John

> I do have that (while I thought it was a good book, I was irked by it because
> I noticed a number of small errors, which was extremely disappointing, because
> he could have had readers who could have caught them, and it mars an otherwise
> excellent work; it makes me wonder if his Jobs bio is the same).
> 
> Yes, his coverage of both of those topics (Kleinrock, and the nuclear war
> meme) is very good, and IMO, correct. Hopefully future writers will follow his
> lead on both of these! :-)
> 
> 	Noel
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.