On Fri, 20 May 2005, Hans Hagen wrote:
Interesting! Found a reference to this also in Forssmann's "Lesetypographie" on p. 81 (new Didot point has been standardized in
ooops, it was the "Detailtypografie" (Forssmann and de Jong).
1978).
ah, that makes the current implementation a bug -)
:-) interesting that Knuth still chose the old definition
btw, similar rounding situations occur with A values:
4*width(A7) <> height(A4)
this one isn't caused by TeX; it's defined by the ISO standard, which tells to calculate A1, A2,... by repeated division from A0, and after each division one has to round to the next lower integer (floor). So all these A* dimensions do _not_ fit together already by the standard (which is IMHO a pity; one can't put a nice sqrt(2) formula on the whiteboard telling that A* paper sizes are strictly following this beautiful principle). Regards, Hartmut