## Forum archive 2000-2006

### Bill Ziemer - Error in problem:

by Arnold Pizer -
Number of replies: 0
 Error in problem: topic started 9/23/2003; 5:28:19 PMlast post 12/31/2003; 5:13:40 PM
 Bill Ziemer - Error in problem:  9/23/2003; 5:28:19 PM (reads: 1032, responses: 5) From the Rochester library setLimitsRates1_5Graphs/ur_lr_1-5_3.pg Seed 1320 gives the incorrect answer. Other seeds are fine. <| Post or View Comments |>

 Coreen Mett - Re: Error in problem:  9/25/2003; 10:04:42 PM (reads: 1187, responses: 0) In about line 182 the evaluator      num_cmp(int( f_rule( g_rule($xg-.00001) )) ) , needs to be changed to num_cmp(rnd( f_rule( g_rule($xg-.00001) )) ) , ------ Now for a question:  How does one make sure that the problem gets corrected in the distribution of *.pg files?   Coreen   <| Post or View Comments |>

 Bill Ziemer - Re: Error in problem:  9/26/2003; 2:41:04 AM (reads: 1189, responses: 0) This exact question, and the need for a solution, is the basis for the construction of a national webwork problem database. Jeff Holt, John Jones, and I have submitted an NSF grant to do this. It should be coming, soon if the grant is funded. <| Post or View Comments |>

 Arnold K. Pizer - Re: Error in problem:  9/26/2003; 8:44:20 AM (reads: 1201, responses: 0) Hi, Hopefully Bill, Jeff, and John's proposal will be funded and we will have a robust national webwork problem database. Until that time the best thing to do is to report bugs here (or to http://webwork3.math.rochester.edu/bugzilla/query.cgi) and we will attempt to fix the bug.  If you want to speed solutions, please send us bug fixes (thanks Coreen). In this particular  case, I have changed "int("    to    "rnd("  in two places (lines 175 and 180 of setLimitsRates1_5Graphs/ur_lr_1-5_3.pg). You can get the corrected version from http://webwork3.math.rochester.edu/cgi-bin/viewcvs.cgi/rochester_problib/setLimitsRates1_5Graphs/?cvsroot=UR+Problem+Library  The corrected version will be included in the next distribution of .pg files. Arnie <| Post or View Comments |>

 Thomas R. Shemanske - Re: Error in problem:  9/26/2003; 3:39:27 PM (reads: 1196, responses: 1) We seem to have another one of the same type, but for which the "int" to "rnd" trick did not work setLimitRates2Limits/ur_lr_2_10.pg with seed 4525 Tom<| Post or View Comments |>

 Thomas R. Shemanske - Re: Error in problem:  12/31/2003; 5:13:40 PM (reads: 1303, responses: 0) I took some time to look at this problem now that the term is over.  There is no error. The given problem is lim_{xto 0^+} f(g(x)) A quick glance at the graph of g show that is goes to zero (through positive values) as x -> 0^+, so one evaluates lim_{zto 0^+) f(z) Sadly, it is the resolution of the graph that causes the problem.  If you *really* stare at in, it appears as though the graph of g dips just below the x-axis right near the origin.  In fact that is the case, so one really has to evaluate lim_{zto 0^-} f(z) for the correct answer Setting the size=>[400,400] options in the graphs allows enough resolution to discern this more easliy. Tom<| Post or View Comments |>