On 6/2/08, William Stein <[EMAIL PROTECTED]> wrote: > > sage: a = e - 1 > > sage: for i in range(2,200): > > ....: a = (a-1)*i > > ....: > > sage: float(a) > > -inf > > <joke> > Oh my god, floating point numbers are just broken! > </joke>
:-) > Seriously, floating point numbers are what they are. Fair enough. The point I tried to make (not claiming anything is broken) is perhaps complemented by the following: sage: expand(a) 3943289336823952517761816069660925311475679888435866316473712666221797249817016714601521420059923119520886060694598194151288213951213185525309633124764149655567314286353816586186984944719612228107258321201270166459320656137141474266387621212037869516201606287027897843301130159520851620311758504293980894611113948118519486873600000000000000000000000000000000000000000000000*e - 10718971748644869545882855035558709219326197160960676962072649602582551264719783234505244404668998392076585003319325446678612269966334945842611842879092400407644778161460862429166937616985702801147176754507686032298026763211111234497467428555982733907104528850484797214679822706672415691688065756374587435109932151999457607946738407969860677357914214979720210208723664522819 This is a particular symbolic expression, which has a particular value in RR, whose best approximation in RDF is far from -inf or sage: RDF(expand(a)) nan The "philosophical" question is whether: sage: RDF(A*e - B) (where A*e is damn close to B+1) should be evaluated as sage: RDF(A)*RDF(e) - RDF(B) which is definitely nan since RDF(A) == RDF(B) == +inf. What I mean is, the symbolic expression has (in principle) enough information to figure out that RDF(A*e-B) is close to 1. Whether it is reasonable to expect that is the question. Best, Gonzalo PS: for the record, this seems to be "broken" in the same way in mathematica, for one. --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
