Split implication and equivalence links into scoping and non scoping types

[ngeiswei]

As discussed in https://groups.google.com/forum/#!searchin/opencog/ImplicationLink$3F%7Csort:relevance/opencog/PI4THQ5etwU/hfMgCpg8AwAJ we want to split ImplicationLink and alike, EquivalenceLink, ExtensionalImplicationLink, etc, into scope and non-scope versions.

I suggest to rename the scope versions

ImplicationScopeLink
ExtensionalImplicationScopeLink
IntensionalImplicationScopeLink
EquivalenceScopeLink
...

I do understand that according to the English language it gives the scope qualifier a stronger position than the implication qualifier, on the other hand it follows the type inheritance prefix order which has been followed so far in the atom type inheritance taxonomy.

[linas]

OK, yes, reviewed, looks good to me. Assuming teh wiki pages have been updated, this can be closed?

[linas]

Hmm. After looking at this more closely, it occurs to me that the following are equivalent:

ImplicationScopeLink
     VariableNode  x
     P(x)
     Q(x)

and

ForAllLink
    VariableNode x
    ImplicationLink
        P(x)
        Q(x)

If these two are NOT equivalent, then how are they different? If they are equivalent, do we really need to have an ImplicationScopeLink?

[ngeiswei]

There are NOT equivalent. Looking for an old email I wrote over a year ago explaining why there are not...

[ngeiswei]

The wiki is documented well enough I believe http://wiki.opencog.org/w/ImplicationScopeLink

I'm closing.

[linas]

Ah, yes. I think Ben said this in an email: that its like "AverageQuantifierLink" or something like that.