Biography:Pat Holmes Sterbenz
Pat Holmes Sterbenz (July 30, 1927 – July 26, 2008) was an American mathematician and computer scientist. He worked at IBM and later was a professor of computer and information science at Brooklyn College. He is best known for his textbook Floating-Point Computation (1974), one of the first systematic treatments of floating-point arithmetic, and for the Sterbenz lemma named after him, which states conditions under which floating-point subtraction is exact.
Life
Sterbenz was born in Ohio on July 30, 1927. He earned a Bachelor of Science degree from Ohio Wesleyan University in Delaware, Ohio, where he met Nancy Norton, whom he married in 1949; the couple had five children.[1] In 1953 he received a PhD in mathematics from Ohio State University with the dissertation Some Topics in Cohomology Theory, written under the supervision of Paul V. Reichelderfer.[1][2]
Sterbenz then joined IBM, where he worked in the emerging field of computer science.[1] At the IBM Systems Research Institute in New York City he taught a course on floating-point computation for several years, out of which his textbook grew.[3] In the early 1970s he left IBM to become a professor of computer and information science at Brooklyn College.[1]
A longtime resident of Larchmont, New York, Sterbenz died on July 26, 2008, after several weeks of hospitalization.[1]
Work
Together with C. T. Fike, Sterbenz wrote two papers on optimal starting values for the computation of square roots by Newton's method.[4][5]
His book Floating-Point Computation[3] presents floating-point arithmetic in a generalized form that allows for variations in radix and word length, and discusses the arithmetic of the IBM System/360 in detail. It contains the result now known as the Sterbenz lemma (Theorem 4.3.1): if y/2 ≤ x ≤ 2y, then the floating-point difference x − y is computed exactly. The lemma remains a standard tool in the error analysis of numerical algorithms under IEEE 754 arithmetic.
Selected publications
- Sterbenz, P. H.; Fike, C. T. (1969). "Optimal starting approximations for Newton's method". Mathematics of Computation 23 (106): 313–318. doi:10.2307/2004425.
- Fike, C. T.; Sterbenz, P. H. (1971). "Minimax approximations subject to a constraint". Mathematics of Computation 25 (114): 295–298. doi:10.2307/2004923.
- Sterbenz, Pat H. (1974). Floating-Point Computation. Prentice-Hall Series in Automatic Computation. Englewood Cliffs, NJ: Prentice-Hall. ISBN 0-13-322495-3. https://archive.org/details/floatingpointcom0000ster.
- Sterbenz, Pat H. (1975). "Understandable arithmetic". 3rd IEEE Symposium on Computer Arithmetic (ARITH). pp. 33–35. doi:10.1109/ARITH.1975.6157004.
References
- ↑ 1.0 1.1 1.2 1.3 1.4 "Pat Sterbenz Obituary". The Journal News. https://www.legacy.com/us/obituaries/lohud/name/pat-sterbenz-obituary?id=48131892.
- ↑ "Pat Sterbenz". https://www.genealogy.math.ndsu.nodak.edu/id.php?id=10170.
- ↑ 3.0 3.1 Sterbenz, Pat H. (1974). Floating-Point Computation. Englewood Cliffs, NJ: Prentice-Hall. ISBN 0-13-322495-3. Reviewed in Rowland, John H. (1976). "Review: Floating-Point Computation". SIAM Review 18 (1): 138–139. doi:10.1137/1018026.
- ↑ Sterbenz, P. H.; Fike, C. T. (1969). "Optimal starting approximations for Newton's method". Mathematics of Computation 23 (106): 313–318. doi:10.2307/2004425.
- ↑ Fike, C. T.; Sterbenz, P. H. (1971). "Minimax approximations subject to a constraint". Mathematics of Computation 25 (114): 295–298. doi:10.2307/2004923.
 |  |
