[Dynamite] OK Guy you win

[Ian Holmes]

(& Ewan too I guess!)

Re local alignments...

I spent a fruitless hour trying to design adapter interfaces to convert
local models into global ones by adding flanking states.

In the end I realised that this truly is the wrong place to impose this
kind of design decision. So I wrote local & global Viterbi algorithm
interfaces.

It needs to be well-defined how these algorithms behave (and I still think
flanking states are the way forward here), but I do now realise that this
is the right way to do it...

so.. have a beer on me

Ian
-- 
Ian Holmes  ....  Howard Hughes Medical Institute  ....  ihh@fruitfly.org