** Using WeBWorK **

Thanks.

Jun

<| Post or View Comments |>

One possibility would be to include PGasu.pl at the top and then use

ANS( pc_evaluator([str_cmp("zero"), 1], [str_cmp("0"), 1], [str_cmp("nothing"), 1], [str_cmp("nothingness"), 1], [str_cmp("zilch"), 0.5]) );

I haven't actually tested the line above, but I think it is right. It should also give 1/2 credit if the student answers zilch.

John

<| Post or View Comments |>

I tried the above code and it gave the following compiling errors:

"Compiler warnings: * Error in PGtranslator.pm::process_answers: Answer AnSwEr1: Not an ARRAY reference at (eval 126) line 275. ##More details: --------Error in PGtranslator.pm::process_answers: Answer AnSwEr1:

---- PGtranslator::process_answers called at processProblem8.pl line 532

-----------------------------------

* Error in PGtranslator.pm::process_answers: Answer AnSwEr1: Answer evaluators must return a hash or an AnswerEvaluator type, not type || at /usr/local/webwork/system/lib//PGtranslator.pm line 931. "

I will update PGasu.pl from your CVS and try it again. Thanks.

Jun

<| Post or View Comments |>

"Error detected while loading /usr/local/webwork/system/courseScripts/PGbasicmacros.pl: PG_restricted_eval detected error at line 259 of file /usr/local/webwork/system/courseScripts/dangerousMacros.pl Subroutine list_random redefined at /usr/local/webwork/system/courseScripts/PGbasicmacros.pl line 801, line 1. The calling package is PG_priv"

Do I need to update the PGbasicmacros.pl?

Jun

<| Post or View Comments |>

First, you may have the wrong version of PGasu.pl from cvs. You should get the one in system/courseScripts . That one should not have a problem with list_random.

Second, I should probably change it to take the syntax I gave above, but the syntax for using it is really:

ANS( pc_evaluator([[str_cmp("zero"), 1, ''], [str_cmp("0"), 1, ''], [str_cmp("nothing"), 1, ''], [str_cmp("nothingness"), 1, ''], [str_cmp("zilch"), 0.5, 'For full credit, try using a less colloquial word']]) );

The third arguments are comments which can be displayed. There is also an extra set of brackets this time.

John

<| Post or View Comments |>

Jun

<| Post or View Comments |>

ANS( pc_evaluator([str_cmp("zero")], [str_cmp("0")], [str_cmp("nothing")], [str_cmp("nothingness")], [str_cmp("zilch"), 0.5]) );

John

<| Post or View Comments |>

Thanks, this is very nice. I think it would be good to have an alias for this called or_evaluator or or_method or something with the word "or" in the name. I assume "pc" stands for positive credit which is reasonable since the method "or's" over the possible answers until it finds one with positive credit but I doubt most people would guess what pc_evaluator does from the name.

I think of this as an "or" method. Jun was thinking of it as allowing "multiple" answers. Can someone come up with a name that suggests what the method does?

Arnie

<| Post or View Comments |>

As I was making modification, I had the same idea. You are right that "pc" is for partial credit (someone here asked for it and then never used it).

Presumably the "or" evaluator could be simpler in its structure, so that

or_evaluator( str_cmp("zero"), str_cmp("nil"), num_cmp(0) )

would just call

pc_evaluator( [str_cmp("zero")], [str_cmp("nil")], [num_cmp(0)] )

I am open to better names than or_evaluator.

John

<| Post or View Comments |>

Also perhaps as a default each answer could be assumed to call the evaluator from the previous answer, unless a new one is supplied. (I haven't looked at the code, I'm just thinking of ways to cut down syntax.):

multiple_eval([str_cmp("zero")],["nil"],["nothing"],[num_cmp(0)])

Mark

<| Post or View Comments |>

multiple_eval([str_cmp("zero")],["nil"],["nothing"],[num_cmp(0)])

is feasible. You probably could have a command which tried all values on the same answer evaluator, like

multiple_eval(~~str_cmp, ["zero","nil","nothing"], optional arguments)

where the square brackets are necessary to separate optional arguments. You would need those in case you wanted to pass extra arguments to the evaluator.

I think

multiple_eval( str_cmp("zero"), str_cmp("nil"), num_cmp(0) )

may be longer, but is more straightforward making it easier for people to remember how to invoke it.

John

<| Post or View Comments |>