In the version of WeBWorK we are running
(checked out by our sysadmin on 16 Sept.
at about 2pm EDT), a student can tell
whether or not their answer is correct by
using 'Preview Answers' (the color in the
answer blanks indicate what is correct and
what is incorrect).
This undermines many matching and
true/false questions completely.
Has anyone else encountered this,
and/or have suggestions for fixing it?
(E.g., is this still the case in the most
current version?)
William Boshuck
Yes -- that bug was fixed last May in revision 6775
(http://webwork.maa.org/viewvc/system/tags/rel-2-5-0/webwork2/lib/WeBWorK/ContentGenerator/Problem.pm?view=log )
If you update to "trunk" or to version rel-2-5-0 you will get this fix.
(or you might be able to just update the file Problem.pm to revision 6775 -- but I can't guarantee results in that case.)
Take care,
Mike
(http://webwork.maa.org/viewvc/system/tags/rel-2-5-0/webwork2/lib/WeBWorK/ContentGenerator/Problem.pm?view=log )
If you update to "trunk" or to version rel-2-5-0 you will get this fix.
(or you might be able to just update the file Problem.pm to revision 6775 -- but I can't guarantee results in that case.)
Take care,
Mike
Hello Mike,
The code is synchronized with 'trunk' (rev. 7047) and Problem.pm is at revision 7038... Is this strange? In any case, I hadn't taken account of this behaviour until William mentioned it so I'm just having a look to see where it's coming from...
Thanks!
Andy
The code is synchronized with 'trunk' (rev. 7047) and Problem.pm is at revision 7038... Is this strange? In any case, I hadn't taken account of this behaviour until William mentioned it so I'm just having a look to see where it's coming from...
Thanks!
Andy
Library/Michigan/Chap1Sec6/Q25.pg modified to have
$showPartialCorrectAnswers = 0;
will not reveal correct/incorrect in the Answer Preview, but it will color the answer blanks.
Including
install_problem_grader(~~&std_problem_grader);
does not alter that coloring.
(Both pg and webwork2 are at revision 7047.)
Looks like this bug crept back in when the .css was updated and the localization added.
I'll put it in bugzilla --hopefully I can figure out exactly what happened and get it fixed tomorrow.
I'll put it in bugzilla --hopefully I can figure out exactly what happened and get it fixed tomorrow.
On the off-chance that it might be related:
In a problem asking for a point as an answer in Context('point'), having both
$showPartialCorrectAnswers = 0;
install_problem_grader(~~&std_problem_grader);
is not sufficient to suppress messages such as
The second coordinate is incorrect
In this case, as discussed on
http://webwork.maa.org/pod/pg_TRUNK/doc/MathObjects/MathObjectsAnswerCheckers.html#flags_for_point____cmp
it is sufficient to use
ANS( $pt -> cmp( showCoordinateHints => 0 ) );
in order to not reveal info about a (partially) incorrect answer.
I suspect the answer coloring situation is different, but I did write "off-chance" ;-)