[e2e] Jon Crowcroft's thought questions...

[Simon Spero]

--On Saturday, July 26, 2003 7:04 AM -0400 Craig Partridge 
<craig at aland.bbn.com> wrote:

>> ii) where did XTP end up

Well, Bert Dempsey had a position at  UNC's School of Information and 
Library Science, but he wasn't doing transport stuff (might startle the 
librarians).
Bert left UNC a few years ago to work for RateIntegration.


> There was also a wonderful private debate c. 1986 about whether the DNS
> approach was right, or whether a replicated central database into which
> you registered would scale better.  As I recall, a lot of the debate
> centered on estimates of the relative cost of bandwidth -- those who
> argued for centralization said, hey the revolution has to come (and it

 This ties in to an interesting counterfactual; during the late 90s bubble 
the telecom sector massively over-invested in bandwidth; as a result, the 
relative cost of bandwidth compared to other factors was seriously 
depressed.  The counterfactual question: how would protocol design have 
been differed if bandwidth had not been underpriced?  Would SOAP exist in 
its current form?

What about if the bandwidth over-supply had occurred in the mid 80s; would 
OSI application protocols have had more success, or were they inherently 
doomed the moment the first session layer draft was printed?

Simon
p.s.

1) Does anybody have a copy of the last HOSTS.TXT file? (or at least, the 
last one that had 10.* hosts in it).  What about a zone file for .arpa?

2) I think I've asked this before, but does anybody have a working Fuzzball 
for sale;if not, what about a free-standing LSI-11 that could be fuzzed up? 
I'd also be interested in a working Lisp Machine.  Yes, it would cheaper 
for me to just buy a new fan heater before the winter, but  it's history, 
dammit :-)