feature request: random seed option

[alexholcombe]

I can see in the code that there actually is a seed option that is mostly implement, it's just not quite there yet. Looks like maybe you ran into a problem during implementation and gave up - I tried to fix it myself but didn't get anywhere.

A working seed parameter would be very useful because (as discussed with @richarddmorey on email) with different randomize runs, one gets different results, with some results better than others, but presently that is not reproducible. It seems that the result depends not on the order that the students that are listed in, but instead the order of the student numbers. Thus the only way I could come up with to run the algorithm repeatedly with different "seeds" was to make a fake student number for the algorithm's input files, and then take the algorithm's output and convert all the student numbers back to the real student numbers. This took a bit of programming, of course.

[richarddmorey]

It is definitely implemented in the javascript code - it's been a while since I've looked at this, but is it not working?

[alexholcombe]

I only know (or am 98% confident, based on me not getting it to work) that it isn't working when called from R.

[richarddmorey]

ok, I'll check it out. From what I remember, it should be passing that parameter into the javascript engine, but maybe it isn't getting picked up fro some reason.