My favorites | Sign in
Project Home Downloads Wiki Issues Source
New issue   Search
for
  Advanced search   Search tips   Subscriptions
Issue 3172: solid geometry -- randomness?
12 people starred this issue and may be notified of changes. Back to list
Status:  Started
Owner:  marcia....@gmail.com


Sign in to add a comment
 
Project Member Reported by marcia....@gmail.com, Apr 1, 2011
http://www.khanacademy.org/exercises?exid=solid_geometry&read_only=1&problem_number=1

http://www.khanacademy.org/exercises?exid=solid_geometry&read_only=1&problem_number=7

http://www.khanacademy.org/exercises?exid=solid_geometry&read_only=1&problem_number=11

Is it pure chance that problem numbers 1, 7, 11 give the same problem? Haven't looked at the code yet, but figured you'd want to check it out first. :) If it is pure chance, then I wonder whether it'd be useful to keep track of what's already been shown and delay repeat problems.

Apr 1, 2011
Project Member #2 d...@desmondbrand.com
Well, that didn't take long :)

I'll take a look now.
Apr 1, 2011
Project Member #3 marcia....@gmail.com
I mentioned that last sentence without giving it too much thought. If it turns out to be a big pain and not worth it (as compared to fixing bugs / new exercises), don't worry about it.
Apr 1, 2011
Project Member #4 d...@desmondbrand.com
So the fact that problem 1 and 7 are the same is a coincidence. But 1 and 11 is not. The prng we are using generates the same sequence when seeded with any of the strings that match /^.+$/

So, that means problem 1, 11, 111, 1111 are all identical. Same with 2, 22, 222, etc. Try it on some other exercises. The first one I tried, prime factorization, has the same behaviour.

I also confirmed it on David Bau's (author of seedrandom.js) site[1]. Go there and type strings into the form. They all generate the same sequence.

I think this this is a bug in seedrandom.js, so I have fixed it there. I posted a comment to David's site asking if he can confirm the fix makes sense.

[1]: http://davidbau.com/archives/2010/01/30/random_seeds_coded_hints_and_quintillions.html
Apr 3, 2011
Project Member #5 d...@desmondbrand.com
David updated seedrandom.js to fix this behaviour for numeric seeds. I've attached an hg export of his changes. Since we are using a string seed, David's fix does not solve our problem, so it shouldn't be absolutely necessary to apply this patch. It might be nice to keep up to date though.

I've attached another couple of changesets that add a prefix to the seed used to initialize the prng. This will make the all exercises have better randomization, however it also breaks all the existing &problem_number=x links. So I added an extra url parameter called bareseed that will prevent the seed from being prefixed. This can be used to reproduce old problem numbers for debugging purposes. Add &bareseed=true to the URL any time you want to reproduce an old problem.


update_seedrandom.export
2.1 KB   View   Download
prefix_problem_seed.export
7.3 KB   View   Download
Status: Started
Cc: marcia....@gmail.com
Apr 4, 2011
Project Member #6 marcia....@gmail.com
Cool, thanks for the quick action. I'll take a proper look later by EOD tmr!
May 3, 2011
Project Member #7 d...@desmondbrand.com
I think this was forgotten :)
Owner: marcia....@gmail.com
Cc: -marcia....@gmail.com
May 4, 2011
Project Member #8 marcia....@gmail.com
 Issue 3756  has been merged into this issue.
Jun 9, 2011
#10 vincent....@googlemail.com
I also experienced the repating of several problems today in the same practice:
Solid geometry. In such a high quantity that I have not yet seen this in other practices.
Apr 7, 2012
#11 rameshku...@gmail.com
Hi

 Under solid geometry, I tried to answer in the units of Tau rather than pi. It did not accept Tau as a unit in place of 2 times pi.
Apr 7, 2012
#12 rameshku...@gmail.com
Hi

 Under solid geometry, I tried to answer in the units of Tau rather than pi. It did not accept Tau as a unit in place of 2 times pi.



Sign in to add a comment

Powered by Google Project Hosting