opencog-dev team mailing list archive

["Joel Pitt" <joel.pitt@xxxxxxxxx>]

On Fri, Aug 1, 2008 at 1:57 PM, Ben Goertzel <ben@xxxxxxxxxxxx> wrote:
> I don't know the exact status of the VH code, I think Welter would be the
> one to ask about that...
>
> My best recollection is that this stuff
>
> https://extranet.vettalabs.com:8443/bin/view/Novamente/AtomVersionHandles
>
> was implemented, but this stuff
>
> https://extranet.vettalabs.com:8443/bin/view/Novamente/ComplexAtomVersionHandles
>
> may not have been.
>
> However, if this is the case, I think that implementing complex
> AtomVersionHandles is of general value for OpenCog, beyond PLN, so it should
> be done by someone fairly soon...

>From my interpretation of the code, it seems complex version handles
are currently not in OpenCog (only a single Handle/indicator is
stored, not a list).

I was wondering what the reason behind storing the indicator
(HYPOTHETICAL or CONTEXTUAL) is? Shouldn't one be able to just look at
the type of the Handle in the VersionHandle and infer this based on
whether the the Handle is a ContextLink or not? ...perhaps it's just
for ease of access, or perhaps it is in case a Handle is deleted but a
TV is still around that involved the deleted handle in an inference.

I guess that, initially, complex version handles won't be necessary
for PLN since each atom will be in it's own temporary dummy PLN
context rather than being nested.

J