Updated 2012-06-10 05:36:11 by RLE

Lars H, 8 Nov 2004: This page was created as an adjunct to Proper integers for Tcl 8.5 or 9.0 to house some code that aimed to implement the ideas there. A more recent and readable (but still incomplete) version of that can be found in [1].

Things completed:

  • Implementation of a bigint Tcl_Obj type.
  • Procedures for + and - using SLM-hexadecimal representation.

Things half-done:

  • Implementation of a bigint subtype SLM-hexadecimal with internal representation.
  • Building using CriTcl

Things hardly even started:

  • Other operations

Talk

I wonder - I recently ran into a reference to http://freshmeat.net/projects/hfloat/ which proports to provide support exteme precision floating point as well. I don't recall hearing anyone ask for huge floating point though; only integers. - RS: Arbitrary-size integers are still of finite length; huge floating-point numbers will need a bound in precision, otherwise any irrational number like atan(-1) (=Pi) will never stop to produce digits...

Sarnold Precision handling in the field of arb. length floats is what is done in BigFloat for Tcl (formerly MPA). The other point is BigFloats should be in decimal form to allow a proper decimal representation, which is not the case in C's doubles.