Home page - David B. Jaffe

David B. Jaffe

Director, Computational Research and Development
Genome Sequencing and Analysis Program
Broad Institute of MIT and Harvard

Most current work: see ALLPATHS-LG.

From 1989 to 2000, I was on the faculty of the UNL Department of Mathematics and Statistics. In 2000, I took leave to work at the Center for Genome Research, and in 2002 I resigned my position as Professor at UNL. I can be reached at jaffe at broadinstitute [DOT] org.

I thank the Department for allowing me to maintain this page. Most of it is completely out of date. In particular, the machines cpthree and bigbox are dead.

Online and maintained, believed to be useful

coding theory database also available in Macedonian
the complete Split source code and detailed results
Split manual (March 21, 2000), in postscript, same as "Binary linear codes: new results on nonexistence, 1100 pages, perhaps more than you want to know about this subject

Weird stuff

Sensory mishmash: me (remove beard), my dead cat, my vita, some ethnic restaurants in Lincoln, opinionated comments about colloquium talks, guide to the computer projection equipment in Oldfather 204.

Not-current Class Schedule and Office Hours (Spring 2000)

Math 208 (Multivariable Calculus) 9:30-10:20 MTWF (OldH 204)
(The class project from last year Math 208 goes to Venus, is still up, for your amusement, but not quite functional.)
Math 939 (Algorithms in Biological Sequence Analysis) 4:30-5:20 MThF (OldH 204). Here was the course announcement.
Old course materials and course announcements

Ancient research Interests - Coding Theory

Before the third millenium, my research area was the existence and classification problem for optimal binary linear codes. For a more detailed description of some problems I studied, see this and the documents shown below. All my results and examples are encoded in a computer language Split which I have designed. If you have viewing or printing problems, look here or write to me.

Funding from the National Science Foundation for all this work (grants DMS-9623205, DMS-9801581) is gratefully acknowledged!

(March 2000) Numerical results on the asymptotic rate of binary codes , joint work with A. Barg. Also you might wish to look at the earlier version, which has graphs in color: Linear programming bounds on codes of length 1000, "transparencies" for DIMACS talk, report on work (which was) in progress, joint with A. Barg.

(July 1999) I just put up the following two papers. Also, the codes database has been updated (and moved). I can also put this database on a CD and mail it to you if you would like (at no charge - it costs me next to nothing), but I am not sure how to port a database which uses cgi scripts.

Optimal binary linear codes of dimension at most seven, to appear in Discrete Mathematics. This is joint work with Iliya Bouyukliev.
The smallest length of eight-dimensional binary linear codes with prescribed minimum distance, to appear in IEEE Transactions on Information Theory. This is joint work with Iliya Bouyukliev and Vesselin Vavrek.

Optimal binary linear codes of length <= 30, to appear in Discrete Mathematics. Updated 11/6/98.
A brief tour of split linear programming, Springer-Verlag Lecture Notes in Computer Science 1255 (1997), 164-173.
Looking inside codes, Proceedings of the Second International Workshop on Optimal Codes and Related Topics (Sozopol, Bulgaria, June 9-15, 1998), Institute of Mathematics and Informatics (Sofia, 1998), pp. 137-143. [Technical note. Use e.g. Cyclic({1,2,3}) now in place of Cyclic(1,2,3).]
Binary linear codes: new results on nonexistence, 1100 pages, version 0.6beta (3/21/2000), ongoing work. This is the main document for Split. See also installation instructions and tips for using Split.
Some of the data from these documents (and their provisional successors) is accessible here, where you will find a form for entering [n,k]; you will get some information about [n,k] codes, including thousands of generator matrices, mostly for codes of dimension <= 12. Last updated 7/23/99.
You can experiment with my program here.
New binary linear codes which are dual transforms of good codes , preprint.
Computing linear codes and unitals, Designs, Codes, and Cryptography 14 (1998), 39-52. This is joint work with Vladimir Tonchev.
A sextic surface cannot have 66 nodes.

Some coding theory links

Even more ancient research Interests: Classical Algebraic Geometry , Picard Groups, Etc.

Ancient links for finding people:

  1. Math Departments: USA, world
  2. Phonebooks: check all US phonebooks (amazing!), Who's Online (math), combinatorics, AMS, who-where
  3. Some people: Laurinda Jaffe, Ken Stanley, Mark Terasaki

Ancient links for making travel plans:

  1. Time (good luck): from Australia
  2. U.S. State Department travel search engine
  3. Everything I know about buying inexpensive airline tickets.
  4. Boston bed and breakfast: phone 1-800-888-0178 or 1-617-576-1492. (Giving e.g. "576-1492" to a search engine is a good way to get related info.)
  5. Airline information: United Airlines, Northwest Airlines (flight info)
  6. Atlapedia country database
  7. You can get to European rail timetables from here. The Netherlands one (NLD) contains a link to an all-European rail computer. (The last time I tried to find this I had trouble, but got ultimately to it.). Here is an index of European railway resources.
  8. currency conversion

Ancient links for libraries and online databases:

  1. UNL Library, ... card catalog, ... keyword search, ... title search, ... author search, ... journal search, ... Interlibrary Loan ( journal article, book).
  2. UNL libraries and beyond: a quick tour from the mathematician's point of view
  3. MathSciNet - Texas mirror (faster for me) or MathSciNet - main site; for both you must have a subscription.
  4. JSTOR (online access to various mathematical journals)
  5. National Library of Medicine search engine
  6. Bibliography of Fossil Vertebrates
  7. Library of congress: homepage, catalog

Ancient software links, Other ancient links