bug in function_cmp_up_to_constant  topic started 4/13/2001; 10:01:46 AM last post 4/16/2001; 8:08:55 AM 





Zbigniew Fiedorowicz  Re: bug in function_cmp_up_to_constant 4/16/2001; 8:08:55 AM (reads: 2101, responses: 0) 
Hi Arnie and Mike, I also found the PGanswerMacros.pl.original version of of function_cmp_up_to_constant() to be vastly more reliable than the current 1.6 version. I tested it on the problem function_cmp_up_to_constant("x^9/36") which was the subject of the student complaint with a variety of seeds. Unlike the current version which accepts such off the wall answers as x^9/20, x^8/30, x^7/36, x^6/36, 1234567 (for most values of randseed), the old version always rejected these and also such comparatively tame wrong answers like x^9/35, x^9/37, etc. It is enough to convince me to switch back to the old version in current classes at OSU. Zig
> >Hi Mike and Zig, > >I just tried this with PGanswerMacros.pl.original (which is the old version >of the macros that we shipped with the new version of WeBWorK) and 123456 >does not work for any problem in Math 142, set 6 except problem 10. I found >the fact that it worked for problem 10 very disconcerting until I looked at >problem 10 which is a definite integral but who ever wrote it used >function_cmp_up_to_constant() for the answer so we can not blame >function_cmp_up_to_constant() for that. > >Zig, try switching to the old version PGanswerMacros.pl in a test course >and see if you still experience problems. > >Arnie > 