[logon] HandOn release schedule

[Francis Bond]

G'day,

> however, i also noticed that PET returns the parse selection score at
> minimal precision (one digit), i.e. nothing after the decimal point.  i
> just corrected that in the `mcheap' binaries (the chart mapping branch
> of PET).  but for the HandOn release i do not want to recompile cheap,
> i.e. the binary built off the main PET branch.  this because there has
> been major re-factoring there lately, and the messages from cj suggest
> that some more debugging may be needed.
>
> hence, francis and berthold, i would suggest you try running with the
> `mcheap' binary for the time being.  without turning on chart mapping,
> we expect it should behave identical to `cheap'.  and it is about time
> we confirm that hypothesis.  you would be paid with four extra digits
> in precision for the parse ranking scores reported to [incr tsdb()].
>
> could you locally edit your `dot.tsdbrc' and try this out?  assuming
> you observe stable behavior, i would then make that change in SVN.

I tried this  just now, and ascore wasn't being stored for mcheap
(perhaps I am too fast :-).
I also note that for jacy it is always null ((:ASCORE)), even though I
load a parsing model.

Back to bed (^_^).



-- 
Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
NICT Language Infrastructure Group