[logon] important bug fix in MRS construction

[Stephan Oepen]

colleagues,

two weeks ago i had inadvertently introduced a bug in the MRS creation
from grammar-internal feature structure, resulting in events being mis-
construed as MRS variables of type `i'.  the bug only affected the old
version of the ERG, or other grammars (and only the SVN `trunk').  dan
and i were busy with a near-complete new release of the ERG, hence did
not notice this issue (until ann found and repaired it).  my apologies
for the hick-up!

the forthcoming ERG version uses the VPM also for mapping of variable
types, i.e. the family of parameters *event-type*, *ref_ind-type*, et
al. becomes unnecessary.  i have updated the VPM documentation:

  http://wiki.delph-in.net/moin/RmrsVpm

i would recommend that both JaCY and GG extend their VPMs to make use
of this extended functionality.  eventually, i would try to deprecate
the traditional mechanism, i.e. *event-type* et al.

another `new' facility is more or less documented on the wiki now:

  http://wiki.delph-in.net/moin/ReppTop

REPP is still work in progress, but used centrally in the forthcoming
ERG release.  i believe only GG has pre-processor rules today, and in
the long run i hope that REPP (plus chart mapping) can replace FSPP.

i will announce these and other developments more widely once we have
completed the ERG release, in a sense our `model solution' for putting
to use all the new toys :-).  but, if nothing else, please update your
LOGON tree to obtain the MRS bug fix mentioned initially.

                                                   best wishes  -  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 ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++