[developers] SEM-I question: properties on 'i' variables

Michael Wayne Goodman goodman.m.w at gmail.com
Mon Nov 12 22:38:51 CET 2018

Hi Stephan,

Sorry if I wasn't clear. I meant that one of those is redundant to the 
other. Since the NUM property is only relevant for 'x' variables, there 
is no practical difference between the two (please correct me if I'm 
wrong). If it were possible for the ARG0 of def_implicit_q to be an 'e' 
variable (i.e., the other subtype of 'i'), then it would not have the 
NUM property at all. So I cannot imagine any situation where the second 
synopsis of def_implicit_q would apply and the first would not also apply.

On 11/12/18 1:28 PM, Stephan Oepen wrote:
> hi mike,
>> Perhaps something in the grammar could be
>> more tightly constrained so the SEM-I generation code doesn't enumerate
>> apparent redundancies such as the following?
>>       def_explicit_q : ARG0 x { NUM sg }, RSTR h, BODY h.
>>       def_implicit_q : ARG0 i { NUM sg }, RSTR h, BODY h.
> why do you think either of the above are redundant?  i agree with you
> (and dan) that it is surprising to see the NUM property on a variable
> of type ‘i’, given the current state of the ‘variables’ section in
> ‘erg.smi’.  but the fact that these two predicates constrain their
> ARG0s to a specific NUM value seems the opposite of redundant to me;
> it is essential information for example to the provider of an input
> semantics for generation and could in principle form the basis of
> rejecting generator inputs upfront where they carry an incompatible
> value (plural, in this case).
> best, oe

More information about the developers mailing list