Arnon, Just wanted to thank you for the white paper "Fallacies of Distributed Computing Explained." I think you did a terrific job putting a lot of good material in a single spot. We are building web applications and the marketplace is laboring under some misconceptions about distributed apps (like it does everything a desktop app does, in the same amount of time, is easier to support and has not environmental conflict issues). Things are not all bad as you point out but many of these misconceptions are based on glossy marketing materials and not the realworld application of the technologies. Thanks for giving us an easy to use, accessible document of the true "lay of the land."

Best Regards.
Kenn

Hi Kenn,
Thanks for the comment and compliments. The real credit should go to Peter Deutsch and James Gosling - I just tried to explain what they said as well as tried to show the fallacies are still valid (and still ignored many times).

Arnon

Seems like we never learn from the past. Check out this SUN paper "A Note on Distributed Computing" from November 1994 at http://research.sun.com/techrep/1994/smli_tr-94-29.pdf

Or to expose the overall strategy (at the cost of obscuring the well-stated & cogent tactics) -

"You know what they say about the word ASSUME!"

:-)

Re: "Latency is zero", I remember in the early 90's developing a Windows-front-end Oracle-back-end client server application. It had low defects in system test and at our first several dozen customers. Then a major customer deployed it with their data center in the US and Asia-Pacific clients. ODBC, a very chatty protocol, was our middleware - with thousands of round-trips per transaction.
We soon learned that latency is NOTHING LIKE zero!