Re: Face header

[Antoine Leca]

address@hidden wrote:
> 
> 1. The standard point, established by international agreement well over 100
> years ago, as used by Knuth and by type foundries everywhere is 1/72.27
> inch

Well, I do not know what is the reason, but for sure French foundries
should be outside "everywhere"... They used, and still use (except one which
is pretty special and except of course for computer typography), the didot
point which is *not* the same as the one that is used in the Britannic
world (that's said, I do not know what use(d) Germans, Italians. Nor I do
care, to tell the truth).


> The 'anglo-saxon point' and 'IN point' are unknown to typesetting practise.

Well, the 'IN point' is. It (intended to) means "le point de l'Imprimerie
nationale", which I can roughly translate to "point of the (French) National
Printing Agency". I know we are going off-topic, but the story is odd enough
that I believe it is worth taling.

During the French "Revolution", as you might know, the government establishes
a new set of units (from which the present SI units), to replace the older
sets of units that happen to vary from place to place (and as we see,
typography was no exception ;-)). So when it comes to typography, request
was made to change from the existing Didot point to a new point, which should
be directly related to the new "metre": the value was chosen to be 0.4 mm,
close enough to minimize practical differences.
And the new agency, created at the same time, was quite obviously urged to
use the new measurement unit; on the other side, the business printing
industrie could not afford the investment, because all the types (matrices)
should be changed by new ones! So as a result, in France we have two
point sizes which were used concurrently since 200 years (to be exact,
before the American software foundries flood up the industry).

Where the story becomes real fun, and that is particularly dedicated
at the people that want to see use of metrical units everywhere ;-),
the guys at the Imprimerie Nationale fumble when it went to calculate
what for a length 0.4 mm was (you should remember that by this time,
there was no ruler with millimeter graduations, obviously). So they
end up with a point that mesured 0.39877 mm (I took Bertrand's value
as correct, I did not checked it), i.e. a 0.3 % error.


As a result, if someone want to use millimeters (a noble cause) for
distances, s/he surely needs to base the precision (of course not
the accuracy) on a much smaller unit, perhaps 10 nm (alias 0.01 µm
or .00001 mm). For instance, Windows that uses a twips which is
1/20 point, so 1.7 µm, falls quite below the mark (in fact, they
fall that far that for some use and in particular in typography,
the unit used causes rounding errors when displayed in millimeters,
and the result is ugly *and* really annoying, like a 1,9999 cm
margin; but that is another problem).


Antoine