
3/3/92

These two are also mentioned below but I think I understand them better now

Constraints with more than one condition on their right hand side
wont match uttsits properly.

    *S : [...]
    <=
       *T : [....],
       *U : [....].

If *U and *T are both utt sits that don't meet (end and start vertices
the same) it doesn't apply.

In a constraint where a variable is mentioned as an argument and also
as a proposition they don't get treated as the same.

27/9/91

Subsituations referred to in constraints as variables with
conditions are not coindex with variables of the same name
Looks like variables matching situations in arguments are not
always bound to situation names in constraints.

MORE 14/2/92 
  <=
    *S : [S ! S != <<fred,S,*T :: [T | T != <<bob,T,1>>],1>>].
doesn't work while
  <=
    *S : [S ! S != <<fred,S,*T :: [T | T != <<bob,*T,1>>],1>>].
does

30/8/91

DONE

   Constraints apply to all situations, grammar rules only to
   utterance situations.  This isn't quite right yet.  Constraints
   referencing mixed sits get their start's and end screwed so
   aren't found in the indexing method

   There's still going to be a problem with rules whose daughter
   conditions require only propositions that are generated by
   normal constraints.  In that case the end vertex wont be bound
   properly which is required for connecting to other utterance
   situations

17/7/91

DONE
   Sub situations in the last conditions fo the last daughter in a rule
   seem to screw up the vertexs so the result can't be used in another
   rule.

   Lexical insertion should existentialise variables.

DONE
   Existentialisation should be done by going through a a term rather
   than collect variables and fixing them.

   Sub-situations in the head on a constraint that also appear 
   in the body aren't treated properly.  There result gets added to
   the condition

------

Known bugs

DONE
    "free" variables in basesits and constraints don't get
    displayed properly

DONE (ish)
    Constraints will loop more that I want

DONE
    Reference to other situations in grammar constraints wont work
    (they wont get the start and end right).

DONE
    Query/parsing hack

    Sub-situations referenced in basesits and lexical entries must
    also be in basesits 

    Allow reduced forms in syntax, (S != <<sings,h,1>>) for
      S : [S ! S != <<sings,h,1>>]

    Give some reasonable error messages in parsing the language

    Constraints always introduce new propositions.  They shouldn't
    do this is one is already there (not sure how to check for that).
But this is right!

Done
    Printing situations should be neater
    
    Prove --> query

    Help and debugging at astl level

