dear all, in a former life, we had a BSCW workspace for PET, where emily once posted a very relevant message.  i am about to shut down the PET workspace, hence forward emily's message below (and will do the same for a few more messages). all best, oe ======= 17-aug-04 posting by emily ======= Hello, The past few months have seen a two minor extensions to tdl as accepted by the LKB: -- Documentation strings, enclosed in " ", in the first line after the parents. (Types only) -- Type addenda statements, signalled by :+ (instead of :=) which allow users to (monotonically) add information to previously defined types: additional parents, additional constraints, and/or additional documentation strings. In BNF, the revisions look like this: Avm-def -> := Conjunction | Comment Conjunction Comment -> " string " Type-addendum -> Type Avm-addendum . Avm-addendum -> :+ Parents Conjunction | Parents Comment <pre> Conjunction | Parents | Parents Comment | Comment Conjunction | Comment | Conjunction</pre> The motivation for the type addendum is from the Grammar Matrix. I hope to soon define modules for, say, basic word order, in which I'll be defining types that I expect grammar writers to want to add to. Defining a subtype to bear the additional constraints is not always an option, and overwriting the existing Matrix type with a new type containing the original info plus the new constraints is a) hacky and b) bad news for distributing Matrix updates. The documentation strings should eventually be browseable within the LKB, and we intend (someday :-) to document the Matrix in this way. The purpose of this note is to request that these extensions be considered for inclusion in PET (flop) as well. Thanks, Emily