Re: Getting greatest decimal accuracy out of G_PI

On Sat, 3 Feb 2007 13:31:57 +0200
Tor Lillqvist <tml iki fi> wrote:

> > Just feeling for the limits of what I can do given the contraints
> > of the language.
>And how is that related to GTK+?

Well Gtk+ is based on Glib. The Glib header defined the accuracy
of G_PI to 50 decimal places.  Isn't it a fair question to enquire whether
the extra functionality provided by GLib, somehow allows us to use it?
I don't see a Glib maillist anywhere.

> > I like to think about things like using the center of the earth as
> > an origin, and using a spherical coordinate system to locate
> > points(and/or areas volumes) on the surface or in space. The accuracy 
> > depends mostly on the accuracy of the angles in radians( ie. the
> > number of decimal places of PI).
>What about the accuracy of the measurements that specify where said
>points are in the first place? What about the accuracy of the
>trigonometric function implementations? What about the accuracy and
>stability (etc, stuff I don't really know) of the algorithms you use?
>You do know, I hope, that for instance (hypothetically) adding a 20
>digit "accurate" 1.0000000000000000000 to an equally accurate
>1.0000000000000000000e-6 a million times isn't necessarily going to
>produce what you perhaps expect.
> > I'm sure it can go into the other direction also, what about nano-technology
> > where you are dealing with the super-small. The ability to distinquish
> > very small angles would be very important.
>Sorry, but I do get the feeling that this is one of the discussions
>where the somewhat blunt but most appropriate answer is "if you have
>to ask that question (on an off-topic mailing list, even), you
>shouldn't be doing it".

Sorry if I bothered you, by raising the topic.


I'm not really a human, but I play one on earth.

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]