Jeffrey D. Ullman

Jeff Ullman is the Stanford W. Ascherman Professor of Computer Science (Emeritus). His interests include database theory, database integration, data mining, and education using the information infrastructure.

What's New | Polemics | Books | Biographical Information

What's New

My son Jonathan Ullman's Web page. Jon is working in the theory of data privacy at Harvard.

Generalizing Map-Reduce

I have been working with Foto Afrati on algorithms that exploit the map-reduce environment (often called distributed file systems) but that are not necessarily constrained by the limitations of Map processes feeding Reduce processes. A Optimizing Joins in a Map-Reduce Environment looks at the multiway join in particular, and shows how to find the optimum way to distribute the work among Reduce processes. It appears in EDBT 2010. A New Computation Model for Cluster Computing is a rewrite of something that was rejected from PODS 2009. We have submitted it to PODS 2010, and we hope the PC will recognize the importance of what we are trying to do: provide a model in which one can talk about the relative merits of algorithms for execution on Hadoop, Clustera, or similar programming systems. (Added later: they again missed the point.)

Book: Foundations of Computer Science

In 1992, Al Aho and I published a book called Foundations of Computer Science, whose goal was to present CS theory as something integrally connected to CS practice. For example, we viewed recursive programs, recursive definitions, and inductive proofs as the same thing. We believe the book deserved a better fate, but the publisher took it out of print years ago. Having received back the rights to the book, we are happy to make it freely available Here.

Gradiance

I'm a founder of Gradiance Corporation, whose goal is to provide better, cheaper homework and programming-lab support for college courses. You can get a Tour of Gradiance. We also have developed automated tutorials in SQL, Java, and GUI Programming.

Gradiance has recently entered into an agreement with Pearson Education (Prentice-Hall and Addison-Wesley) to offer on-line homeworks in connection with many of the Pearson books in Computer Science and Engineering. The material for the Garcia-Ullman-Widom books is available now, but we are working on materials for a number of books in Databases, CS1/Java Automata, Compilers, Operating Systems, and Data Mining. In North America, Europe, Africa, and the Middle East, you should contact your Prentice-Hall or Addison-Wesley representative. Outside those areas, you can arrange for services directly from Gradiance (email to support at gradiance dt com) or through Pearson.

I Would Like to Hear From You, But...

I generally enjoy getting emails, even if it is to tell me of a mistake in a book. In fact, those notes are particularly important; they help not only me and my coauthors, but more importantly, later readers of the book. I try to respond helpfully and in a timely manner to email that isn't spam. However, there are two classes of emails that I think should not be written and that get a form response. Read More.

Polemics

As part of my role as old curmudgeon (replacing my previous role as young curmudgeon), I have been writing a few articles about things I think especially stupid. I hope to write more.

Answers to All Questions Iranian. Am I the only one who gets peppered with questions from Iranians? They're technical questions or political questions, but there seems to be a system behind them. So I decided to save time and answer them all here.
Fundamentalism. Shortly after 9/11/01, I wrote this article on fundamentalism --- the nonnegotiable belief in some unproven hypothesis. It's been added to periodically, and rambles a bit, over the foolishness of the Palestinian leadership, the drug warriors, the politically correct, the anti-abortion crowd, and a few others. This article also advocated research on a system like "Total Information Awareness," although I doubt that it had any influence on that development.
4-Way Stop Signs. Dumb traffic policies have always been way up on my list of things I'd like to fix but can't. The global problem is that those with the power to decide, focus on specific risks, which they can see. But in solving one problem, they often introduce far more harmful effects --- e.g., pollution. The problem is that the harm is distributed sufficiently widely so that no one can point to a particular death and attribute it to a particular instance of some dumb traffic policy. This polemic, originally written for Club Nexus (the prototype for Orkut), focuses on the 4-way stop signs on Stanford Campus, but the principles apply broadly.
Attack of the Fifty Foot NIMBY's. Throughout my life, I've been infuriated by the tendency of our social structure to allow theft, as long as the loss is distributed in tiny parcels among a large number of people. "Affirmative action" is probably the best known of these subtle thefts, but this essay is not about affirmative action. Rather, it's about a theft that occurs on our streets every day, yet no one notices.
Software Patents. This article was written for the Knuth Prize. Don, who is of course ineligible for the Knuth Prize, would probably say we should get rid of all patents. I don't agree, but I do feel that in the software area, things have gotten out of hand. Here, I tried to argue that courts should apply the same standards that are used when choosing papers worthy of presentation at a conference, to the matter of whether a certain software idea deserves a patent.
University of California (Non)Confidentiality Policy. The UC system, unlike every other university in this country, refuses to offer the maximum confidentiality that the law allows for letters of recommendation that they solicit. Members of the various computer-science departments in the UC system understand that this policy makes letter-writing irrelevant and deprives them of an important source of advice. However, the policy comes from so high up that it is impossible for them to do anything about it. I have therefore endeavored to put in place a solution similar to that used in response to the equally foolish "Buckley amendment," which gave student applicants the right to see their recommendations. The document referred to is a pledge that the person about whom you are writing a letter can sign, to assure you that you can write an honest and frank letter without fear of later reprisal or embarrassment. I wish the UC departments would themselves solicit this pledge, in order to preserve your anonymity as well as your privacy, but apparently they are forbidden to do so.

Books --- Past and Future

For some sets of notes and materials for supplementing current books, click here:

Biographical Information

Jeffrey D. Ullman
ullman @ cs.stanford.edu
650-494-8016 (home)
650-725-2588 (FAX)