WeBWorK Problems

How are problem seeds generated?

by Patti Lamm -
Number of replies: 3
I recently noticed that if I look at a student's assignment data prior to a homework set's opening date, the seeds for the problems on that assignment seem to be randomly given: 1021, 38, 998, etc.

But after the opening date (or after the student first clicks on the assignment?), the seeds change to something like 998, 998, 998, 998, ..., for every problem on the same assignment.

The strange part (to me) was that this same seed appeared on TWO different assignments which opened on the same day.

It would be helpful to understand what webwork is doing behind the scenes. For example, is it likely that if I take a problem from an assignment and put that same problem on a gateway quiz, will a student have the SAME SEED in both cases?

Thanks,
-Patti Lamm
In reply to Patti Lamm

Re: How are problem seeds generated?

by Gavin LaRose -
Hi Patti,

My short answer is "I don't think this should happen." I just tested it on my system, creating a set with two problems, assigned to me. The seeds for the two problems start as 258 and 2944, and don't change. Unassigning and reassigning it to myself, the seeds change to other values, but again don't change thereafter.

Is this for a normal homework assignment? Does it happen for all students? Which problems are you using in the assignment?

Gavin

In reply to Gavin LaRose

Re: How are problem seeds generated?

by Patti Lamm -
Hi Gavin,

It's a normal homework assignment and it happens for all students. Not sure what you mean by "which problems" are in the assignments -- these are problems we've written, and this behavior happens for all assignments in all seven courses using webwork.

When the sets are newly assigned, the seeds (for a two-problem set) would be something like you say -- 258 and 2944. But as soon as the set opens, they will change for student #1 to be something like 35 and 35. I suspect that unassigning and reassigning would lead to a similar behavior but for new seeds initially, and then new (repeated) seeds afterward.

I only saw one situation where a student had identical (repeated) seeds for two assignments, and these were assignments which opened at the same time.

Sound like a database problem?

We're running 2.4.9.

Thanks,
-Patti
In reply to Gavin LaRose

Re: How are problem seeds generated?

by Patti Lamm -
OK, so I think we've figured it out. The values of the seeds aren't changing according to when the set opens, as I previously thought. Evidently it's related to scripts that we're running to assign homework sets to students. Any homework set that is assigned manually (using the webwork interface) gives random values of the seeds. Any homework set that is assigned using a script appears to have the problem of repeated values of seeds for an individual student on a given assignment.

Sorry for the posting -- but thanks for letting me know that it's not a webwork phenomenon. Now to look more closely at our scripts...

-Patti