On Wed, Apr 11, 2007 at 12:48:31PM -0400, Stavros Macrakis wrote:
> It seems unlikely that a higher number of default digits (20 rather than 16)
> would actually be useful to anyone. Just reducing the default fpprintprec
> would probably be enough. But it would be possible to have floating-point
> notation be interpreted as bigfloats by default if necessary (I've already
> written the code...)
I find that in my calculations I am rarely in need of more than say 6
decimal digits in a final answer. These are generally calculations
involving things like fitting curves to measurements, and solving
differential equations to model various engineering issues... In
applied situations it is extremely rare for a measurement apparatus to
be available with 6 digits of precision. Even when they are (such as
some electronic mass balances) I tend to doubt their accuracy. The
variation in earth's gravitational constant from location to location
on the surface of the earth is I believe easily within the range of a
balance capable of weighting 1 gram +- 1e-6. Since these balances are
based on gravitational forces even if they read .070306 g the actual
mass of the object in question might be +- .00005 g without you
knowing it
In some fields, such as surveying, large precisions are used. In a
surveying class I took we located a certain point to within .0001
survey-foot (.030480061 mm) on our drawings. This point was determined
by a concrete nail with a head approximately 5mm in diameter!
When I pointed out that the typical amplitude of a sound wave
travelling through concrete might involve vibrations that would vary
this number significantly this issue was swept under the rug quickly
by the professor.
I think it would be fairly safe to default the fpprintprec to
something between 6 and 8. Those needing more display precision will
be motivated to figure out how to get it.
--
Daniel Lakeland
dlakelan at street-artists.org
http://www.street-artists.org/~dlakelan