From: SourceForge.net <no...@so...> - 2003-08-15 03:23:56
|
Bugs item #789048, was opened at 2003-08-14 23:00 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104933&aid=789048&group_id=4933 Category: None Group: None Status: Open Resolution: None Priority: 5 Submitted By: Stavros Macrakis (macrakis) Assigned to: Nobody/Anonymous (nobody) Summary: allroots issues (doc and precision) Initial Comment: The documentation doesn't say to what precision allroots calculates its argument: in fact, it gives full floating-point precision (even if the arguments are rats or bfloats). allroots is advertised as "find[ing] all the ...roots of the real polynomial..." In fact it works on real AND complex polynomials, but requires that the coefficients all be explicit numbers (which is not stated). The following get errors: allroots(x-sqrt(2)) => ALLROOTS: not a polynomial allroots(x-%pi) => ** error while printing ERROR message ** ALLROOTS: polynomial not univariate: ~M Shouldn't allroots simply float its argument? allroots((x - 1)^10) gets answers about 20% off because it multiplies out the expression before looking for roots. Shouldn't already-factored expressions be left that way? (I am not sure what the right answer is to that, by the way -- perhaps it is more uniform to expand them all as floating-point polynomials?) In any case, this should be documented. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104933&aid=789048&group_id=4933 |