I am at funconf in Ireland. After a ride in a Delorean, a private train ride to Galway and a helicopter flight I am sitting outside a cottage on the island of Inishmore. The attendants are an amazing group of people, mostly engineers, and I am humbled to be invited along. Everybody is equally passionate about "doing the right" in tech, in business and in life. The discussions shift quickly form one to the other.
I brought two papers with me on this part of the trip, one was the paper that Ben Black sent out earlier this week on Data-Structures for Geometric Approximation, which is a great in-depth review of that area and it will take me a while to finish it.
But I am sitting here outside the cottage on a stone wall reflecting on the other paper I brought. It is another Jim Gray classic. What is amazing is that so much of the work that Jim and Bruce Lindsay and others did at the end of the 60's and into the 70's at IBM and Tandem was so truly fundamental to distributed systems today. They are our heros. 25 years later when Jim was my mentor, he displayed the same unbounded positive enthusiasm for new ideas that he must have had in those early days. The good thing is that in many of my fellow funconf peers I see the same unconstraint energy for wanting to solve the hardest problems. The future looks bright.
Today's reading are Jim Gray's thoughts on lock managements and consistency from 1976. Jim had already started working on locking at the end of the 60's so this paper is a reasonable summary of his thoughts.
Granularity of Locks and Degrees of Consistency, J. Gray, R. Lorie, G.F. Putzolu, and I.L. Traiger, Modeling in Data Base Management Systems, G.M. Nijssen ed., North Holland Pub., 1976, pp. 364-394.