Message-Id: <9210011713.AA23632@sc.zib-berlin.dbp.de>
Reply-To: Mailing list for the LaTeX3 project <LATEX-L@vm.urz.Uni-Heidelberg.de>
Date:         Thu, 1 Oct 92 17:57:26 +0100
From: Paul Taylor <pt@DOC.IC.AC.UK>
Sender: Mailing list for the LaTeX3 project <LATEX-L@vm.urz.Uni-Heidelberg.de>
To: Multiple recipients of <LATEX-L@vm.urz.Uni-Heidelberg.de>
Subject:      Re: rotation of boxes using \specials
Status: R


> The DVI driver standards committee has
> proposed a rotation  special which has been "on the books" for
> several years now. The syntax is \special{rotate N} where N is
> the angle in 1/65536 degrees.

Why on earth do you have to measure rotation in degrees? That means that
if you want to draw a diagonal line between two (essentially cartesian)
given points you have first to calculate the inverse tangent (in TeX).
Arithmetic in TeX is bl**dy awful, and arctan is probably the most difficult
trig function to calculate.

Suggest instead (as I've used in my commutative diagrams package) a ratio,
as used by LaTeX but, of course, without the restrictions. Let the
PostScript interpreter calculate the arctan.

I do (using Rokicki's \specials):
        \special{ps: x y bturn}
        \hbox{}
        \special{ps:eturn}
where
    /bturn {%                           % x y axisht on PS stack
        gsave                           % save graphics state
        currentpoint                    % push position
        currentpoint translate          % move origin there
        4 2 roll neg exch atan rotate   % rotate baseline
                % dvips doesn't know we've changed the coordinate system, so
        neg exch neg exch translate     % move origin back
        } def
    /eturn {currentpoint grestore moveto} def

I also do my arithmetic using continued fractions.

Paul Taylor
Dept of Computing
Imperial College
London SW7 2BZ, UK
+44 71 589 5111 x 5057
pt@doc.ic.ac.uk

PS commutative diagrams package available by anonymous ftp from
theory.doc.ic.ac.uk 146.169.2.37
/tex/contrib/Taylor/tex/diagrams-V4*