[developers] Whoa, missed someting, sorry

Ann Copestake Ann.Copestake at cl.cam.ac.uk
Tue Mar 25 18:55:13 CET 2008

kiefer at dfki.de said:
> What exactly are "the same options"?

There are a number of parameters which can be changed.  From what Yi says, it 
should be fine - I hope that at some point I'll get a chance to look at the 
PET behaviour but can't see that being soon.

kiefer at dfki.de said:
> So what exactly is necessary for the MRS->RMRS conversion? 

I hope that's reasonably clear from the papers/reports (the Lisp code to do it 
is reasonably well commented).  One thing - as far as the description of RMRS 
goes, I suggest looking at the paper in last year's deep grammar workshop at 
ACL.  The conversion from MRS is quite a lot simpler for the version of RMRS 
that does not have INGs, which is the one I would advocate these days and is 
described in that paper.

> Which one
> contains more information, or are they not comparable?

It is always possible to convert MRS -> RMRS, given suitable feature/value 
mappings.  Vice versa is only possible for a subset of RMRSs.

kiefer at dfki.de said:
> At the moment, we have to keep in synch with the changing versions of  ecl
> and the numerous problems that must be circumvented. And as long as  we have
> only one (or maybe two) formats to support, this should be  doable. The rest
> would (hopefully) just be converters from this format,  which could be
> shared?

how were you thinking they would be shared?  currently there is Lisp code to 
do the various formats, so if someone writes a convertor, the Lisp code would 
have to call that if we share the convertor?  I'm not yet convinced this would 
make the overall maintenance load easier ...



More information about the developers mailing list