Discussion at Tomar on how we can further extend the Lexical Type database with other resources for looking up results (led by Francis and scribed by Bec).

FCB: LTDB intro pseudo xml in grammar to document, with examples from treebanks

from Montse: why only lextype?

FCB: no reason, coming what else to do? link to RDF store, or fangorn, or? local is good for robustness, but not as flexible

add type hierarchy to RDF to be queried? can query poss_rel or potential for ex

respond to queries on list like all japanese irregular verbs, raise delphin profile

look for double prefix (un - re- x)

what is best example? shortest? by corpus?

is a common query page/service useful? what else do people want?

bec: type lookup already possible, by url

oe: fcb ambitious, rdf capable of representing semi, could be predicate hierarchy too, coming RSN, can sort of in fingerprint language

wp: wild star surface string doesn't quite get it

oe: clickable fingerprint in ESD, integrate with LTDB, need an agreement RDF store will be up/stable/maintained, fangorn too? need to consider inside/outside view

eb: integrate treebanks into descriptive grammar, want the inside view

tim: fangorn is inactively developed

eb: mix and matching syntax and semantics in query

tim: example sentences - protypical for dictionary is different from prototypical to grammar, perhaps needs more research

fcb: i agree

fcb: also link to lex semantics (WN) or senses

eb: how to set up ltdb for a grammar?

fcb: easy, mostly documented, i can help. will do it for mandarin soon. possibly add type to matrix types. who uses LTDB?

tim, francis, dan, bec, woodley, oe

wp: the full forest treebank links to LTDB

tim: typediff->fangorn also links to LTDB and something else

oe: query re hpsg test suite, csli test suite. sent link (erg.delph-in.net) to items in treebank. pieces of [incr() tsdb] available from web demo

fcb: could link LTDB to same place, rather than static

oe: got web traffic

fcb: sg government blocks something necessary to get there

fcb: encourage grammarians to set it up

dan: ran script on new version erg - works fine?