Re: Possible source of beam properties confusion

[Andrew Bernard]

Hi David,

Thanks so much for this. That particular MWE does indeed print neatly
backward to the order, but precisely as you say, I was seeing random
ordering in my much larger set of examples. Being completely unaware
of this situation that you have here so clearly described, I was
having trouble coding, as the values I was seeing in my debug output
seemed to vary in unintelligible ways, due to my totally naive
assumption that the ordering was linear and the same as the score. Not
so. So now I know.

Your help very much appreciated.

Andrew



On 16 July 2016 at 02:21, David Kastrup wrote:

> The order in which callbacks are called is pretty much arbitrary.