Mathematics Written in Sand - : The hp-15C, Intel 8087, Etc
Mathematics Written in Sand - : The hp-15C, Intel 8087, Etc
Mathematics Written in Sand - : The hp-15C, Intel 8087, Etc
1983
W. Kahan,
University of California @ Berkeley
__________________________________________________________________
hp-21,25,35,45,55,65 10 127. *
Commodore SR4148R 12 127. *
hp-80 Financial 10 13.
TI Business Analyst, SR-30,40 11 100.
Commodore SR4190,5190 12 12.
Commodore SR1400, TI-MBA 12 0/0 Error
TI SR-52,56,51-II 12-13 128.
TI SR-50,50A,51,51A,58,58C,59 13 14.
Monroe 326 13 12.
Enter data:
n := 60*60*24*365 = 31,536,000 sec. per year.
i := 10/n = 0.000 000 317 097 9198 % per sec.
PV := 0
PMT := -0.01 = one cent per sec. to the bank.
·-----------------+-----+--
------------+-----------------------------> x
------------· | |
N x0 x1
[∫ ]
x
Therefore the key must be as fallible as
y
all other sampling procedures. Spikes or jumps
or violent oscillations can precipitate failure.
y
↑
y=2 ----- U=π ---------+--------- U=π --------
: :
y=1 ----- U=0 ---------+ :
| :
y=0 - - - +--------- U=0 ----------> x
!
x=0
___________________________________________________________________________________________________________
File MathSand.pdf
Mathematics Written in Sand Version of 22 Nov. 1983
REFERENCES:
[1] D. B. DeLury "Computation with Approximate Numbers" The
Mathematics Teacher 51 (1958) pp.521-530.
[2] W. Kahan "A Survey of Error Analysis" in "Info. Processing 71"
(1972) pp.1214-1239; North-Holland Publ. Co., Amsterdam.
[3] W. Kahan & B. N. Parlett "Can You Count on your Calculator?"
translated by W. Frangen into "Können Sie sich auf Ihren
Rechner verlassen?" in "Jahrbuch Überblicke Mathematik 1978"
ss. 199-216; Bibliographisches Institut AG , Mannheim.
[4] W. Kahan "Interval Options in the Proposed IEEE Floating Point
Arithmetic Standard" in "Interval Mathematics" ed. by K.
Nickel (1980) pp.99-128; Academic Press, New York.
[5] W. Kahan "Why do we need a floating-point arithmetic standard?"
in preparation.
[6] The Appendix "Accuracy of Numerical Calculations", pp.96-211 in
reference [15] below.
[7] R. E. Martin "Printing Financial Calculator Sets New Standards
for Accuracy and Capability" Hewlett- Packard Journal 29 #2
(Oct. 1977) pp.22-28.
[8] D. W. Harms "Improved Algorithms: Making 23 = 8 " in Session
32 "Advanced Pocket Calculators" of the IEEE ELECTRO 76 in
Boston, May 11-14 1976 . An extract appears in the Hewlett-
Packard Journal 28 #3 (Nov.1976) pp.16-17 .
[9] W. Kahan "Personal Calculator Has Key to Solve Any Equation
f(x) = 0 " Hewlett-Packard J’l 30 #12 (Dec. 1979) pp.20-26.
[10] W. Kahan "Handheld Calculator Evaluates Integrals"
Hewlett-Packard J’l 31 #8 (Aug. 1981) pp.23-32.
[11] Mary H. Payne "Floating Point Basics and Techniques" in the
proceedings of the Spring DECUS Symposium held in St. Louis
May 23-27, 1983.
[12] M. H. Payne and R. N. Hanek "Radian Reduction for Trigonometric
Functions" SIGNUM Newsletter 16 (Jan.1983) pp. 19-24, and a
sequel in Newsletter 17 (Apr. 1983) pp.18-19 .
[13] H. G. Diamond "Stability of Rounded Off Inverses Under
Iteration" Math. of Comp. 32 (1978) pp. 227-32.
[14] P. Henrici "Essentials of Numerical Analysis (with Pocket
Calculator Demonstrations)", and its "Solutions Manual"
(1982); Wiley & Sons, New York.
[15] "HP-15C Advanced Functions Handbook" (1982) part # 00015-90011;
Hewlett-Packard, Corvallis, Oregon.
CONTENTS:
Page Subject
1 ABSTRACT
2 INTRODUCTION
3 VisiCalc 1.10 anomaly
5 WHO’S TO BLAME
7 Degree 504 polynomial
9 Perverse subtraction
10 THE AREA OF A TRIANGLE
12 Theorem: Exact subtraction
13 Diff: a cure worse than the disease
14 FINANCIAL CALCULATORS
15 A Penny for your Thoughts
17 Yield from a Risky Investment
17 SOLVING EQUATIONS
18 Lemma & Theorem about Newton’s Iteration
20 Secant Iteration Phenomenon
21 HP-15C [SOLVE] Key
22 THE [INTEGRATE] KEY
23 Example: the Error function
25 Can’t be foolproof
26 THE CALCULATOR OWNER’S HANDBOOK
27 Counsel of too-near perfection
29 Inescapable anomaly
30 Tolerable noise
32 The Intimidation Factor
32 COMPLEX NUMBERS AND MATRICES
34 A Slab, a Strip, a Channel; Table 1
37 Robust formula for arccosh
38 Matrix examples
40 THE INTEL i8087
43 THE PROPOSED IEEE STANDARD
44 Floating-point exceptions
46 ACKNOWLEDGMENTS
46 REFERENCES