Similarity sorting with the current A*B/C^2 puzzles hangs the flash game

Scenario: go to, click on a design, and try to sort by simillarity.  The flash then becomes unresponsive.  I’ve confirmed with Eli that this is not unique to my machine.

@nando, FWIW

  • it doesn’t seem to be a browser issue (occurs in Chrome, Safari and Firefox).
  • there’s no error indication in the debug console
  • there’s no noticeable increase in CPU activity
  • sorting of the same labs from round 1 still works as expected.
    My guess is that we’ve run into a new issue with the number of designs in one puzzle, but I haven’t taken any action to confirm or falsify that hypothesis.

If this actually does work for anyone, please post that, so we can narrow down the root cause.


Something weird must have happened to eternacac when he submitted the solution 7385370: it’s sequence field was null. This was “killing” any attempt to sort the list by sequence. I deleted the solution.


Thanks for the quick response, Nando.  Could you check the DEC version, too?  It has the same issue.

Looks like jandersonlee submitted 4 designs which are one base short. That too will screw the sorting function. I’m hesitating whether to “patch” them with a final Adenine, or just delete them… Maybe he could fix that himself the way he’d prefer. Is he easily reachable?

feel free to delete them. (I failed at that.)

I see Jeff gave you his answer.  But personally, if there is no significant difference in the implementation effort, it seems that padding it with one or more A’s will almost always be more helpful than throwing the sequence away.  

Well, only saw jandersonlee’s reply, so I deleted them. I’m pretty sure he can re-submit proper versions of these.
Sort function seems restored.

Ah.  I only spoke up because, for whatever reason, I was thinking you were contemplating a code change to avoid the problem in the future.

I don’t think I want to code something here, because any such fixes (in the applet) wouldn’t help actually correcting the problem for the collection process where sequences have to be validated as well. I’d rather keep the code simple and be alerted in case of problems.

The web code could be slightly improved, but I suspect that the first case you reported was caused by something abnormal right at the moment of interaction with Drupal and/or the database, and the invalid record would still have been there, despite checks and safeguards in both the applet and the backend.