[RPRWG] Dave James
Dave James was discovered dead Tuesday, of intentional carbon
monoxide asphyxiation. He was under treatment for serious depression.
Depression causes people to do unreasonable, irrational things;
treatment is not always successful.
Dave was my closest friend.
Over the 20 years I knew him and worked with him, I found him to be
absolutely honest and forthright, and a true professional--he always
strove for the most general and correct design possible within the
time and budget available and other practical considerations.
He was the most ego-free person I've ever known--he would put forth
the best solutions he could see, but if someone showed him he'd made
an error, or there was a better solution, he immediately abandoned
his own brainchild in favor of the better one, and did his best to
give credit to whomever made that contribution. He was extremely
patient with ignorance or honest doubters; less so with BS artists,
bullies, or those who intentionally obfuscate, complexify, or
sabotage for their own personal or corporate gain. This generated
enemies who caused him great pain, as well as depriving us all of
some of his best work. One could argue that his approach to life was
unwise, even foolish in a political sense, personally very costly.
But it was always ethical.
Dave believed that until ideas or designs are described clearly and
unambiguously in written and graphic form, and in considerable
detail, it was not possible to evaluate them. As a result, he worked
hard to document his own work and that of others (even, perhaps
especially, his opponents) in a professional way that exposed
unsolved problems whenever possible. He could generate good
documentation about six times faster than any other person I've
known! It was amazing to see. His generosity was unfortunately not
I first got to know Dave well when he gradually became the chief
architect for the Scalable Coherent Interface project, ISO/IEC
13961:2000 (the original IEEE 1596 (1992) has still not incorporated
the (few) corrections that were put into the ISO/IEC document). This
was a very ambitious project, which thanks in great measure to Dave's
experience and efforts completed in about 2 years, achieving far more
than I'd seen done in any 10 years of my previous experience. It was
far ahead of its time, and was used mainly in supercomputers instead
of becoming the standard PC chip I/O/memory/MP interconnect we'd
hoped for. (It (as IEEE 1596.3) was also the origin of the ubiquitous
LVDS signals that are in wide use today, essentially the only legacy
of this work that is visible to the public.) He also made major
contributions to Futurebus, SerialBus (FireWire), a unifying software
architecture for these (IEEE 1212-1991), and many other standards, as
well as to his employers' projects at several companies.
When I first encountered Dave, at the start of the SCI project, I was
very skeptical--I didn't think it likely that he was correct about
several of the things he seemed to think he knew. Gradually I became
convinced, as I watched him argue the issues with some of the top
people in the field, and usually convince them. When they convinced
him instead, it amazed me how he could change course rapidly to take
advantage of his new understanding, and work with them to reach even
better solutions. It was a real privilege to see this process in
People like Dave don't come along often. I'm glad he came along
during my life. I'm going to miss him a lot.
p.s. Dave's family intends to hold a celebration of life after a
couple of months have passed.
Condolences can be sent to:
Mrs. Barbara James, Silja, and Spencer
3180 South Court
Palo Alto, CA 94306
--David B. Gustavson, tel:650/961-0305 fax:208/475-7525
1946 Fallen Leaf Lane, Los Altos, CA 94024-7206 dbg@SCIzzL.com