Slashdot Mirror
Michael Stonebraker Wins Turing Award
An anonymous reader writes: Michael Stonebraker, an MIT researcher who has revolutionized the field of database management systems and founded multiple successful database companies, has won the Association for Computing Machinery's $1 million A.M. Turing Award, often referred to as "the Nobel Prize of computing." In his previous work at the University of California at Berkeley, Stonebraker developed two of his most influential systems, Ingres and Postgres (PDF), which provide the foundational ideas — and, in many cases, specific source code — that spawned several contemporary database products, including IBM's Informix and EMC's Greenplum. Ingres was one of the first relational databases, which provide a more organized way to store multiple kinds of entities – and which now serve as the industry standard for business storage. Postgres, meanwhile, integrated Ingres' ideas with object-oriented programming, enabling users to natively map objects and their attributes into databases. This new notion of "object-relational" databases could be used to represent and manipulate complex data, like computer-aided design, geospatial data, and time series.
He deserves it. Amazing how many "unknown" people are so important to modern life, since they are usually never talked about.
For anyone involved in databases, Mike Stonebraker is not an unknown, he is a giant. I participated in some of his seminars in Berkeley some years ago. He is a very friendly guy, and a great instructor. He was able to explain deep concepts in simple language, and was patient with dumb students like me. I am happy to see him win this award.
Postgres is the predecessor of PostgreSQL. Postgres used a different query language when it was still a university project led by Stonebraker. Postgres was the next project after Ingres as the name suggests, and its query language was originally similar...called QUEL instead os SQL or something like that.
Postgres forked into two code bases after the university project ended. Stonebraker started a company called Illustra to sell a commercial version of Postgres. Informix eventually bought Illustra and called it Online Dynamic Server if i recall, and by the time IBM bought informix this sibling of Postgres was the flagship product.
The second fork of postgres was picked up by former students of Stonebraker (initially Joly Chen if i recall and one or two others..too lazy to google for the details). They introduced a SQL parser front end of their own and called the initial release Postgres95 v1.x since it was the fad MSFT started to use years in product names, and also resetting the version number given the changes in features and management (postgres was at version 4.x).
When it came time to release the next major version the name was looking dated and redundant since there was still a release number. So the name became PostgreSQL as it was more meaningful (the primary feature difference being the query language). The version number was then "un-reset" too...postgres95 1.x being considered as 5.x and the first PostgreSQL named release being 6.0.
So yes, postgres isn't the same as postgresql. It is mist accurately described as the father of PostgreSQL and Informix. Architecturally the latter two are essentially the same, but their SQL parsers are unrelated as they were each developed post fork, plus the codebases diverged quite significantly over the past 20 years.
They are both fantastic databases by the way...they wipe the floor with mysql. To say postgreSQL is not web scale are ignorant and probably last used it in the 1990s if at all. It truly kicks butt for full text search, geospatial data for mapping or survey data, astronomy and so on. It is 10+ years ahead of Microsoft SQL server or mysql at that stuff as well as things like multi version concurrency...i was spoiled by PostgreSQL MVCC when i had to contend with rows and tables being locked until transactions wete committed in other RDBMSes.
MySQL has no extensibility, nowhere near the rich set of data types or extensibility, and is not optimised for write heavy ACID transaction stuff. MySQL is great for your CD collection or your blog or whatever, but PostgreSQL is still far superior for accounting/erp/mapping/etc, though i do acknowledge MySQL/MariaDB has gotten "good enough" it is far frpm the best.
And dont start with me on noSQL. Its a great hammer but only some applications are nails, even at "web scale".
Congratulations to Dr Stonebraker. His legacy in the industry is impressive and his work has led to a Free database project that can truly take on the big O on many serious fronts.