Changelog for AERN-Real-

If this renders badly as markdown, see the plain text version 28 July 2009: fixed a few bugs in meta-data 0.10.0: 28 July 2009 * switching to beta status * new QuickCheck test suite covering most functionality * new support for anti-consistent intervals (eg [2,0]) (also called directed or improper; using Kaucher arithmetic) * new support for inner-rounded interval arithemtic * fixed errors in some elementary functions for extreme values * fixed performance bug in arctan * improved hierarchy of auxiliary modules

0.9.9: 23 February 2009 * Small changes needed in other AERN packages: * New operation for domain boxes: get its dimension. * Exponentiation, sine, cosine and arctan signinificantly improved for arguments further away from 0. * Fixed a bug in sine Taylor series error term. * Some interval arithmetic operations now have also "inner" versions that approximate the maximal extension of the operation from inside (useful for testing the normal "outer" versions).

0.9.8: 1 December 2008 * added instance of the HTML class for intervals * added syntactic comparison of variable-indexed domain boxes * some extra miscellaneous functions * moved miscellaneous facilities for STM from AERN-RnToRm-Plot to here so that they can be used by AERN-Net 7 October 2008 * hmpfr interface now uses a faster toDouble conversion 30 September 2008 * switched the Demo program to a more suitable (ie faster) base

0.9.7: 30 September 2008 * made it easier to switch among various bases (double, mpfr, pure haskell floats...) * added MPFR backend via hmpfr (cabal install -f "use-hmpfr") * added two new samples (computing pi, inverting Hilbert matrix) that demonstrate the speedup when using MPFR 7 August 2008 * revamped package description to make it much shorter and linked it to the main module

0.9.6: 7 August 2008 * improved domain box class interface and implementation * fixed broken domain box splitting function * improved the integer logarithm auxiliary function

0.9.5: 24 July 2008 * new operation for testing disjoing interiors * real approximations not automatically instances of Ord because comparison is not decidable in general; one should use the four-valued compareReals instead of <, =<, == * removed rependency on haskell98

0.9.4: 15 July 2008 * fixed buggy formatting of floating point numbers * added a simple although inefficient linear solver 12 July 2008 * fixed email in cabal maintainer field

0.9.3: 12 July 2008 * Fixed Data.Number.ER.Real so that it is usable as a single import for the library and its documentation links are more useful. * Added a module with some tests, which can also serve as an example. * Improved formatting of floating point numbers.

0.9.2: 11 July 2008 * declared dependency on haskell98 in cabal file (thanks to Don Stewart)

0.9.1: 11 July 2008 * fixed licence specification within modules

0.9.0: 11 July 2008 * initial release of AERN-Real