[logon] preparing for another LOGON snapshot: Barcelona

Stephan Oepen oe at ifi.uio.no
Wed Jun 17 14:48:08 CEST 2009


> > as mentioned on the DELPH-IN `developers' list today, i would suggest
> > that we jointly produce another tagged LOGON snapshot, i.e. a release
> > of the complete tree than can serve as a stable point of reference for
> > users.  the last such snapshot (HandOn) is getting out of date, and i
> > would like to aim for at least one release each year, synchronized to
> > the annual summit.
> 
> Great idea.   I will try to have an up-to-date  Jacy, JaEn, maybe EnJa.
> 
> I will probably also have KoJa and JaKo as well as an EnEn paraphrase
> system ready by then. Are we aiming to be an inclusive release?  If so
> I will bundle them up too.

well, we are always aiming to be inclusive, but given that both you and
i are a little constrained in our time these weeks, we should calibrate
our ambitions carefully :-).

assuming KRG can now be used without proprietary third-party software,
i would be happy including it in the LOGON tree.  should i just make a
copy of the KRG SVN trunk, and call that `khu/krg/'?  i could then make
available directories `uio/koja/' and `uio/jako/', and let you populate
those with transfer grammars.  i assume you have MT configuration files
too, and i could add these to SVN and create cpu definitions.  however,
if we decide to add these, i think they should be reasonably polished,
i.e. we should not be embarassed if anyone where to look at these files
(or worse, use them as role models for additional language pairs).  if
the best we can hope to do at this point were to drop mildly half-baked
versions into SVN, then i think i would rather leave that to after the
Barcelona release.  i.e. these new language pairs will then be part of
the LOGON `trunk' (which, a few months from now, we will recommend for
people to use anyway, as Barcelona will be out of date), until the next
complete release.  what do you prefer, francis?

> Probable:
>  * LexTypeDB

for that to be included, the code should go into the LKB `trunk' before
thursday next week.  26-jun is the last time i plan to pick up changes
to LKB code for this cycle.  is that a realistic goal?

>  * Useful unknown word handling

not sure what you meant by that?  i realize i owe everyone a follow-up
on the earlier thread on the `developers' list, but we have made some
changes in the ERG already (which dan and i hope are useful :-), and i
believe that is all we can expect to do for this release; the world is
not perfect, but in this case it is at least better than it used to be.

                                                       cheers  -  oe

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++     CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++       --- oe at ifi.uio.no; oe at csli.stanford.edu; stephan at oepen.net ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++



More information about the logon mailing list