The Colloquium Series of the
            Department of Computer Science,
		  University of Wyoming

			presents

                       Jeremy Siek
              University of Colorado, Boulder

  "Languages and Libraries for High-Performance Computing"
 
                  Thursday, March 8, 2007
                         EN 4066
                     4:00 - 5:00 p.m.

Abstract:

Modern computational problems require software systems that are both
complex and efficient. Abstraction is an important tool for managing
complexity but the tension between abstraction and efficiency is a
long-standing difficulty in computer science. Generic programming is
an emerging methodology and tool set that delivers both abstraction
and efficiency. Generic programming has enabled the development of
highly-efficient libraries in key areas of scientific computing such
as numerical linear algebra, network algorithms, and bioinformatics.

In this talk I address issues of modularity that arise in the practice
of generic programming. In particular, generic programming in C++
relies on the template feature, but C++ does not provide a mechanism
to specify the interface of a template. I present a type system and
interface description language that enables modular checking of
template definitions and template uses and that is expressive enough
to handle sophisticated generic libraries. I conclude with a
discussion of ongoing work on the C++ Standards Committee to add this
extension to the next revision of the C++ Standard.

Biography:

Jeremy Siek is a Visiting Assistant Professor in the Department of
Computer Science at the University of Colorado at Boulder and a
Research Scientist at LogicBlox, Inc. His research interests are in
programming languages, generic programming, and
high-performance software libraries. Jeremy is the author of the Boost
Graph Library and a member of the ANSI/ISO C++ Standards
Committee. Jeremy earned a Ph.D. at Indiana University and a B.S. and
M.S. at the University of Notre Dame.