Hi Doug,
I might be mistaken but I *thought* in the past I'd seen Lon-Capa
recognize different ways of expressing the same value as "you've
entered that answer before" and not counting the attempt in
numericalresponse problems.
Submitting (incorrect) answers of 0.650 and 0.65 for a
numericalresponse item results in tries being charged for both
submissions in LON-CAPA versions 2.9, 2.10, and 2.11.
Checking if the same answer was submitted for a numericalresponse item
is determined in the end_numericalresponse() routine in
/home/httpd/lib/perl/Apache/caparesponse.pm, which in turn calls the
check_for_previous() routine in /home/httpd/lib/perl/Apache/response.pm
The check involves a string comparison of the current submission with
each of the previous submissions made by the student for the same
instance of the response item.
If there is an exact match between the value of the current submission
and the value of a past submission, (and the awards also matched),
then the message: 'You have entered that answer before' will be shown
and no try will be charged.
(Exception: survey items, or items where the random seed has changed,
i.e., randomizetry parameter in effect).
That string comparison part of the implementation has not changed from
response.pm rev. 1.32 (8/3/2001) and caparesponse.pm rev. 1.38
(8/3/2001).
Stuart Raeburn
LON-CAPA Academic Consortium
Quoting "Mills, Douglas G" <dmi...@illinois.edu>:
Hi All,
I might be mistaken but I *thought* in the past I'd seen Lon-Capa
recognize different ways of expressing the same value as "you've
entered that answer before" and not counting the attempt in
numericalresponse problems. Am I wrong about that? I thought
entries such as .650 and 0.650 were recognized as the same answer
and the second entry then not counted as a try where there a limited
number of tries, but if that had been the case in the past, it does
not seem to be the case now. A student reported using her two
tries on a test retake with just those two answers (why, I don't
know!) and my subsequent testing of that and some other items
confirmed that attempts are not counted by values but by different
expressions of those values -- as if the problem were a
stringresponse. That was not what I expected. Thank you for any
further guidance on that. Could it be related to changes addressing
differences between decimals with trailing zeros (so .650 is not
the same as .65 for purposes of sig figs) or something? Or was I
imagining that those different expressions were recognized as the
same value in the past?
Doug
Douglas Mills
Director of Instructional Technologies
Department of Chemistry
University of Illinois
dmi...@illinois.edu<mailto:dmi...@illinois.edu>
(217) 244-5739
_______________________________________________
LON-CAPA-users mailing list
LON-CAPA-users@mail.lon-capa.org
http://mail.lon-capa.org/mailman/listinfo/lon-capa-users