[tex-k] Re: wrong paper sizes in config.ps

[Thorsten Schwander]

 > Werner LEMBERG wl at gnu.org
 > Fri Mar 14 12:55:50 CET 2003
 > 
 > It seems that the B series of paper sizes is plain wrong in config.ps.
 > 
 > For example, B4 should be 250mm x 353mm, but config.ps contains
 > 257mm x 364mm.
 > 
 > It's a shame that nobody has ever noticed that earlier...
 > 
 > 
 >     Werner


there are 2 different standards for the 'B' series at work and this should be
decided with some better background information


According to

ISO/DIN B5 paper is 176x250 mm^2 == 499x709 postscript points
JIS (Japan Industrial Standard) it is 182x257mm^2 == 516x729 ps points

JIS A series sizes A0 through A10 are identical to ISO/DIN A series sizes,
JIS B series sizes are not identical to ISO/DIN B series sizes. In fact JIS B
sizes are a tad over 6% larger than ISO/DIN B sizes. While ISO/DIN B sizes
can be cut from JIS B stock (with waste of about 5.7%), JIS B sizes cannot be
cut economically from ISO/DIN B stock.



I am not saying that the JIS numbers should be used, I am just saying that
someone at some point adopted the JIS standard and changing it now may have
some unexpected side effects and should be carefully considered.




As a sample side effect of this confusion, the simple latex file

\documentclass[b5paper]{article}
\usepackage{hyperref}
\begin{document}
does this break gv for you?
\end{document}

results in a ps with a BoundingBox that makes gv (3.5.8) on redhat linux with
any gs 6-8 segfault. Dooh!!!

%!PS-Adobe-2.0
%%Creator: dvips(k) 5.92a Copyright 2002 Radical Eye Software
%%Title: bla.dvi
%%Pages: 1
%%PageOrder: Ascend
%%BoundingBox: 0 0 499 709
%%EndComments
...

In the end I still consider this a flaw in gv (shouldn't segfault from this).

Cheers
T.