On Tue, May 22, 2018 at 12:34 AM, Nil Geisweiller <[email protected]>
wrote:


>
>> What do you suppose GetValueLink to do?
>>
>
> For instance
>
> GetValueLink
>   <atom>
>   <key>
>
> would return the value corresponding to `key` in `atom`. Note that
> `key` is itself an atom. However the returned value may not
> necessarily be an atom, it may be a proto atom, if so it would need to
> "atomized". This makes me think that ProtoAtoms probably need some
> "atomize" method or something.
>

Yeah. OK-ish, maybe, sort-of-ish.  There's a bunch of ugly little details
that is going to torpedo this proposal.

I would be much much happier if, instead, we had a collection of predicates
that responded to questions like "is it true that this value has some
property?" -- for example, "is it true that object X is to the left of
object Y?"  This would be a lot easier than trying to shove x,y,z values
into NumberNodes, and then doing some kind of numerical computations inside
the atomspace.

I  strongly urge that we look at some practical examples, first.

>
>
> It seems it would be good if Atomese offers operators to create,
> destroy, etc, atomspace,


This is coming, its part of gituhub issue #something-or-other.  An
important pre-requisite is clearing out most SetLink usages.  Once this is
done, then we can create an AtomSpaceLink that would be a lot like a
MemberLink. (conceptually. performance requires a rather strange
implementation.)


> because then temporary computations could be
> moved in small atomspace that I suspect would still be able to
> manipulate atoms with less overhead (though I don't know well from
> where come the overhead of inserting atoms in an atomspace).
>

The primary overhead is the scanning of an atom, to see if it is unique
(viz, if we already have it  in the atomspace, or not) This is not
conceptually hard, but it places an upper limit on the atom insertion and
deletion rate. Its the primary bottleneck.

This scanning takes slightly longer in large atomspaces, but is still
significant for tiny atomspaces.  As a rule, you want to avoid placing
NumberNodes into the atomspace.

Linas.

-- 
cassette tapes - analog TV - film cameras - you

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/opencog.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/CAHrUA37akO1AUZXC4MZDp_c-t1%2Bx9xQadrrevq6xwfT4o6Yzuw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to