multi-term extended-precision floating point arithmetic with dbl<n>

[jfcg]

Greetings,

dbl<n> is a compact & fast C++ template library for extended-precision
floating point arithmetic. it is currently beta software, but if want
to try it out and tell me what you think, it is available at http://jiyusoft.com

Thanks,
Serhat

note: multi-term fp representation are described in the paper
"Adaptive Precision Floating-Point Arithmetic and Fast Robust
Geometric Predicates"
by J. R. Shewchuk