From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.io/gmane.comp.tex.context/10598 Path: main.gmane.org!not-for-mail From: Guy Worthington Newsgroups: gmane.comp.tex.context Subject: Re: Euler math again (virtually) Date: 29 Jan 2003 19:51:01 +0800 Sender: ntg-context-admin@ntg.nl Message-ID: References: <20030129014054.16478@mail.comp.lancs.ac.uk> Reply-To: ntg-context@ntg.nl NNTP-Posting-Host: main.gmane.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Trace: main.gmane.org 1043840647 5699 80.91.224.249 (29 Jan 2003 11:44:07 GMT) X-Complaints-To: usenet@main.gmane.org NNTP-Posting-Date: Wed, 29 Jan 2003 11:44:07 +0000 (UTC) Return-path: Original-Received: from ref.vet.uu.nl ([131.211.172.13] helo=ref.ntg.nl) by main.gmane.org with esmtp (Exim 3.35 #1 (Debian)) id 18dqdP-0001TW-00 for ; Wed, 29 Jan 2003 12:44:03 +0100 Original-Received: from ref.ntg.nl (localhost.localdomain [127.0.0.1]) by ref.ntg.nl (Postfix) with ESMTP id 9D11310AF6; Wed, 29 Jan 2003 12:45:47 +0100 (MET) Original-Received: from main.gmane.org (main.gmane.org [80.91.224.249]) by ref.ntg.nl (Postfix) with ESMTP id A81C410AF5 for ; Wed, 29 Jan 2003 12:44:32 +0100 (MET) Original-Received: from list by main.gmane.org with local (Exim 3.35 #1 (Debian)) id 18dqc9-0001Ob-00 for ; Wed, 29 Jan 2003 12:42:45 +0100 X-Injected-Via-Gmane: http://gmane.org/ Original-To: ntg-context@ref.ntg.nl Original-Received: from news by main.gmane.org with local (Exim 3.35 #1 (Debian)) id 18dqc7-0001Ny-00 for ; Wed, 29 Jan 2003 12:42:43 +0100 Original-Lines: 183 Original-X-Complaints-To: usenet@main.gmane.org User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2 Errors-To: ntg-context-admin@ntg.nl X-BeenThere: ntg-context@ntg.nl X-Mailman-Version: 2.0.13 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: mailing list for ConTeXt users List-Unsubscribe: , List-Archive: Xref: main.gmane.org gmane.comp.tex.context:10598 X-Report-Spam: http://spam.gmane.org/gmane.comp.tex.context:10598 Adam Lindsay wrote: > I was wondering if anyone else was interested in the Euler math font. I > was a bit frustrated in its incomplete coverage of certain symbols (e.g. > delimiters, radicals), but then I discovered the Virtual Euler Math > fonts/eulervm for LaTeX. Sounds great. > Does anyone do enough math typesetting to want to test "zeuler" out? > I've gotten good enough results this evening, but I don't have a lot > of sophisticated formulae lying around to do a full test. Does anyone > have some semi- demanding (non-AMS-territory) math for me to try? Here's a script used to test the mathtimes fonts: (courtesy of YandY). I simply commented out the bits that were giving ConTeXt indigestion. %-tryCHIRONMT.ctx----------------------------------------------- \starttext \input CHIRONMT \stoptext %--------------------------------------------------------------- %-CHIRONMT.TEX-------------------------------------------------- %output=pdf % This is a test file containing both math and text. Run in plain TeX %\nopagenumbers % Make sure that `encode.tex' is set up for the encoding used by % your DVI driver for the Times-Roman text fonts. %\input texnansi % deal with TeX 'n ANSI encoding in text % \input ansiacce % deal with Windows ANSI encoding in text % \input stanacce % deal with Standard Encoding in text fonts %\input mtplain % load Times-Roman macros %% NOTE: also change the definition of \qtr (ring accent) if encoding changed! % \ifnum\the\ss=251\def\qtr#1{{\rm\mathaccent202#1}}\fi % standard (Adobe SE) % \ifnum\the\ss=223\def\qtr#1{{\rm\mathaccent176#1}}\fi % ansinew (Windows ANSI) % \ifnum\the\ss=25\def\qtr#1{{\rm\mathaccent23#1}}\fi % textext (TeX text) % \ifnum\the\ss=255\def\qtr#1{{\rm\mathaccent6#1}}\fi % tex256 (Cork DC) % \ifnum\the\ss=167\def\qtr#1{{\rm\mathaccent251#1}}\fi % mac (standard roman) % \ifnum\the\ss=222\def\qtr#1{{\rm\mathaccent176#1}}\fi % texannew % \def\qtr#1{{\rm\vec #1}} % \def\qtr#1{\vec #1} \def\qtr#1{{\rm\mathaccent23#1}} % for TeX 'n ANSI encoding \def\vct#1{{\bf #1}} % vector (bold) \def\uvct#1{{\bf\hat#1}} % unit vector (bold and hat) \def\bvct#1{{\bf\overline#1}} % barred vector (bold and overlined) \def\mat#1{{\bf#1}} % perhaps a bit too heavy for matrix? \def\qand{\quad{\rm and}\quad} % \quad AND \quad \def\qqand{\qquad{\rm and}\qquad} % \quad\quad AND \quad\quad \newdimen\bthick % thickness of lines used in constructing stencils \bthick=0.48pt % 2 pixels at 300 dpi \def\boxit#1{\vbox{\hrule height\bthick\hbox{\vrule width\bthick\kern6pt \vbox{\kern6pt#1\kern6pt}\kern6pt\vrule width\bthick}\hrule height\bthick}} % \def\boldify#1{\hbox{\rlap{$#1$}\kern .6pt{$#1$}}} % moby kludge! \def\sumi{\sum_{i=1}^n} \def\sumiw{\sumi w_i} \def\qq{\qtr{q}} \def\qd{\qtr{d}} \def\ql{\qtr{\ell}} \def\qr{\qtr{r}} \def\qzero{0} % \def\qzero{\qtr{0}} \def\qa{\qtr{a}} \def\qb{\qtr{b}} \def\qs{\qtr{s}} \def\qt{\qtr{t}} \def\qe{\qtr{e}} % \def\tc{\vec c} \def\tc{c} \def\dqq{\delta\qq} \def\dqd{\delta\qd} % \def\vl{\boldify{\ell}} \def\vl{\ell} \def\vr{\vct{r}} \def\vb{\vct{b}} \def\vc{\vct{c}} \def\vd{\vct{d}} \def\vq{\vct{q}} \def\vf{\vct{f}} \def\vg{\vct{g}} \def\vh{\vct{h}} \def\vx{\vct{x}} \def\vy{\vct{y}} \def\dlambda{\delta\lambda} \def\dvx{\delta\vx} \def\jac{{d\vh\over d\vx}} %\centerline{\twelvebf Symmetry in the Coplanarity Condition} \vskip .1in \noindent We can rewrite the triple product in $\qr$, $\qd$, and $\ql$ using $$t = \qr\qd\cdot\qq\ql=\qr\cdot\qq\ql\qd^*=\qq^*\qr\cdot\ql\qd^*.\eqno{(1)}$$ Noting that $\ql^*=-\ql$ and $\qr^*=-\qr$, since $\qr$ and $\ql$ are quaternions with zero scalar parts, % we can rewrite the coplanarity condition in the form we obtain, perhaps surprisingly, $$\boxit{\hbox{$\displaystyle{ t = \qr\qq\cdot\qd\ql }$}}\eqno{(2)}$$ The symmetry between $\qq$ and $\qd$ can perhaps be seen in more detail if the dot-product for $t$ is expanded out % Now expand the dot-product for $t$ in terms of the scalar and vector components of $\qq=(q,\vq)$ and $\qd=(d,\vd)$: $$ t = (\vd\cdot\vr)\,(\vq\cdot\vl) + (\vq\cdot\vr)\,(\vd\cdot\vl) + (dq - \vd\cdot\vq)\,(\vl\cdot\vr) + d\,[\vr\ \vq\ \vl] + q\,[\vr\ \vd\ \vl].\eqno{(3)}$$ At this point we remember that $$\qs = \sumiw e_i \,(\qr_i\qd\ql_i^*) \qand % \quad {\rm and} \quad \qt = \sumiw e_i \,(\qr_i^*\qq\ql_i).\eqno{(4)}$$ % We also still have the three constraint equations $$\qq\cdot\dqq = 0, \quad \qd\cdot\dqd = 0, \qand \qq\cdot\dqd+\qd\cdot\dqq=0,\eqno{(5)}$$ % all of which we can combine in the matrix form % $$\pmatrix{ A & B & \qq & \qzero & \qd \cr B^T & C & \qzero & \qd & \qq \cr \qq^T & \qzero^T & 0 & 0 & 0 \cr \qzero^T & \qd^T & 0 & 0 & 0 \cr \qd^T & \qq^T & 0 & 0 & 0 \cr} \pmatrix{\dqd \cr \dqq \cr \lambda \cr \mu \cr \nu \cr} = -\pmatrix{\qs \cr \qt \cr 0 \cr 0 \cr 0 \cr},\eqno{(6)}$$ Overall, we have a system of 11 equations in 11 unknowns, four of which are the components of $\qq$, four are the components of $\qd$, and three are Lagrangian multipliers. % \noindent Note that the upper left $8\times8$ sub-matrix is the weighted sum of dyadic products $$\sumiw \tc_i\tc_i{}^T,\eqno{(6)}$$ where the eight-component vector $\tc_i$ is given by $$\tc_i = \pmatrix{ \qr_i\qd \ql_i^* \cr \qr_i^*\qq\ql_i \cr} = -\pmatrix{\qr_i\qq\ql_i \cr \qr_i\strut\qd \ql_i \cr}.\eqno{(7)}$$ % We conclude that the special system of equations has the number of solutions that is equal to the number of ways of partitioning the set of variables, in the indicated manner, namely $${n+m-2\choose n-1}= {n+m-2\choose m-1}= {(n+m-2)!\over(n-1)!\,(m-1)!}\eqno{(8)}$$ This typically is {\it much\/} less than the number of solutions of a general homogeneous system of $(n+m-2)$ second degree equations namely $2^{n+m-2}$. The continuation method involves taking a small step $\dlambda$ in $\lambda$ and solving for the increment $\dvx$ in $${d\vh\over d\lambda}\,\dlambda + \jac\,\dvx =0,\eqno{(9)}$$ where $J=(d\vh/d\vx)$ is the Jacobian of $\vh$ with respect to $\vx$. The updated solutions $\vx'=\vx+\dvx$ will not be exact if we are taking finite steps, so one needs to use Newton's method to improve their accuracy. \end %-------------------------------------------------------------- It works for me using cm* fonts.