[epe-users] re-release of preprocessed parser inputs

[Stephan Oepen]

colleagues,

one team alerted us to a serious error in our preprocessed parser
inputs: character counting was reset at sentence boundaries, which of
course gave invalid stand-off pointers.

we have just released an update to the parser inputs, where the ‘raw’
texts remain unchanged (since Version 1.1), and the preprocessed ‘.tt’
files have been corrected:

http://epe.nlpl.eu/index.php?page=4

we have received a couple of trial submissions already, but we would
very much appreciate receiving additional parser outputs in the next
few days—to further debug the task set-up.

we are currently debating whether to share a running score board of
end-to-end results on trial submissions with all participants, or
whether to merely communicate trial results one-to-one.

best wishes, oe