- From: <jhd@csd.uwo.ca>
- Date: Thu, 9 Mar 2000 19:29:32 -0500
- To: www-math@w3.org
- Cc: jhd@maths.bath.ac.uk

\documentclass{article} \def\arccot{\mathop{\rm arccot}\nolimits} \begin{document} \title{Comments on MathML 2\\ Version 2U306 --- Stephen Watt's copy\\ edition of 9 March 2000} \author{James Davenport \tt jhd@maths.bath.ac.uk\\ jhd@csd.uwo.ca} \maketitle Stephen very kindly lent me his copy: here are my observations. \begin{description} \item[p. 113; Line 10 of 4.1.2.] ``BA-level or Baccealaureate''. Would it be sensible to add ``etc.'', or name, say ``Abitur''? \item[p. 116; line 5]This says that the default is that a {\tt ci} comes from a commutative field. Apart from the pedantic point that all fields are commutative, is this really necessary? After all, variables often stand for integers. Maybe all you ned is a commutative ring. I also note that page 125, at line 2 after {\tt ci} says ``no default is specified''. \item[p. 120]Ath the bottom, it is stated that the definition of inverse trigs can differ sligtly. A stronger statement is true: the definition of $\arccot(-1)$. \begin{tabular}{llrl} \cite{AS}&1st printing&$\frac{3\pi}4$&inconsistent\\ \cite{AS}&9th printing&$\frac{-\pi}4$\\ \cite{CRC}&30th edition&$\frac{3\pi}4$&inconsistent\\ Maple&V release 5&$\frac{3\pi}4$\\ Axiom&2.1&$\frac{3\pi}4$\\ Mathematica&\cite{Mma}&$\frac{-\pi}4$\\ Reduce&3.4.1&$\frac{-\pi}4$&in floating point\\ Matlab&5.3.0&$\frac{-\pi}4$&in floating point\\ \end{tabular} \item[p. 121; top]No doubt the definition of {\tt inverse} has already been fought over, but mightn't it be better to say ``for some $x$ in $D$''. Otherwise, since $\arcsin(\sin(3\pi))\ne3\pi$ (normally), one might have problems saying what $\arcsin(-1)$ is. \item[p. 121; 13 lines up]What is ``the scope of the declare''? See similar unspecified phrases on p. 128. \item[p. 126]It is not stated that {\tt type=normal} is the default for {\tt set}: I assume that it is. \item[p. 126; {\tt set and list} +9]``the order defaults to a numeric or lexicographic ordering''. Does this mean that $(z,y,x)$ renders as $(x,y,z)$?, or does this only apply to lists generated via constructors? \item[p. 271]This described {\tt and} as $n$-ary, but {\tt or} and {\tt xor} as binary. This is certainly inconsistent, and is inconsistent with p. 129, which declares all three to be $n$-ary --- DPC says that p. 129 is consistent with the DTD \begin{verbatim} <!ENTITY % clogicopnary '%MathML.and.qname; | %MathML.or.qname; | %MathML.xor.qname;' > \end{verbatim} OpenMath defines all three as $n$-ary, which seems like the correct answer. \item [p. 338; third point in chapter 4]. ``Classifical''. \end{description} More comments to follow, \begin{thebibliography}{9} \bibitem{AS} Abramowitz,M. \& Stegun,I., Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables. US Government Printing Office, 1964. 10th Printing December 1972. \bibitem{Mma} Wolfram,S., The Mathematica Book. Wolfram Media/C.U.P., 1999. \bibitem{CRC} Zwillinger,D. (ed.), CRC Standard Mathematical Tables and Formulae. 30th. ed., CRC Press, Boca Raton, 1996. \end{thebibliography} \end{document}

Received on Thursday, 9 March 2000 19:29:45 UTC