In a couple of recent messages about tagger evaluation, ENGCG and
EngCG-2 were mentioned. Since these taggers were sometimes
inaccurately identified with a number of different institutions and
companies, I thought it might be a good idea to shed some light on the
actual origins of the systems.
The first version of ENGCG was developed at the University of
Helsinki, at Research Unit for Computational Linguistics. In the early
1990s, an early version (ENGCG or ENGCG-1, if you like) was licenced
by its authors via Lingsoft for commercial purposes. To the best of my
knowledge, Lingsoft has not further developed ENGCG.
Further development of the tagger continued at U Helsinki, and
recently in Conexor oy. The resulting tagger, EngCG-2, is currently
available from Conexor (for a demo, see http://www.conexor.fi/analysers.html).
Regards,
Atro Voutilainen