ARPANET versus (#progressinCS)

I’ve been doing a little computer history research. It seems that ARPAnet was contracted for in 1969 and launched on schedule later that year (Wikipedia). This page says that the budget was $2.2 million ($14 million in 2015 mini-dollars). I will leave it to readers to compare to…

6 thoughts on “ARPANET versus (#progressinCS)

  1. And think what was involved. To launch a website you call a library that encapsulates DNS/SSL/HTTP for you. To launch ARPANET … first you have to invent TCP/IP…

  2. ARPAnet predates TCP/IP by a decade and half. The original scope was modest: allow ARPA-funded researchers from any university to log on remotely to ARPA-funded computers at any other university, i.e. the Internet was originally built for the ancestor of rlogin/telnet/SSH.

    The idea behind ARPA was to find brilliant people, give them oodles of money with minimal oversight or accountability, and let them do their thing. Remember, the agency was launched in the aftermath of the Sputnik crisis. Another example is the Manhattan Project, which delivered the nuclear bomb for the equivalent of $30B.

    That said, just like the occasional enlightened monarch (e.g. Lee Kuan-Yew) does not make a case for despotism, episodic flashes of government hyper-competence like ARPA do not make it for unaccountable government either.

    The suddenness of ARPA’s establishment played in its favor – apparatchiks did not have the time to colonize it at first. That happened later. The history of the CIA, rooted in Yale’s Skull & Bones (with some mentoring from the British) is the same – it takes a while for the rot to set in, but the Iron Law of Bureaucracy is inexorable.

    That said, inventing the Internet is a fairly straightforward engineering project compared to developing a system that can conform to tens of thousands of pages of legislation that emerge from the lobbyist-led compromises, and thus expand to hundreds of thousands of pages of specs that are absolutely unfathomable to any mind.

  3. Creating ARPAnet was a shiny fun challenge for some of the smartest CS minds on the planet at the time.….was not. Had the creation of the ARPAnet been handed over to a government contractor (Lockheed, TRW, Boeing, etc.) on a cost-plus basis, the outcome probably would have been similar to

  4. I think Phil is just comparing the cost vs. benefit ratio of the two projects and how absurd things have gotten lately.

    Most endeavors these days involve lot of wasted time and money on just management. The layers of middle management are becoming more pervasive as things are made more complex (via regulation, legislation, etc). It’s the reason why hours spent in the office are increasing for most workers – if management were efficient there would be no need for long hours. Poor management requires an always “on-call” staff to help make up for deficiencies in planning (more reactive than proactive).

  5. @J. Peterson

    ARPAnet was actually built by a government contractor, Bolt Beranek and Newman (BBN), founded by MIT professors, and originally focusing on acoustic engineering. They designed the Interface Message Processors (IMP), essentially routers. There are pockets of excellence among government contractors, e.g. Lockheed Skunk Works, RAND.Corporation

    That said, the modern TCP/IP based Internet owes as much to the French Cyclades project as to ARPAnet.

Comments are closed.