On Fri, Feb 2, 2018 at 7:28 AM, Ben Goertzel <b...@goertzel.org> wrote:

> We are undertaking now to port MOSES into the Atomspace (specifically
> to make each MOSES deme an Atomspace, while leaving the
> deme-management, feature-selection etc. code intact...)...
>

I'm not sure who "we" are, in this sentence.  The reason that "we", meaning
"me"
did not try this in the past is that the Atomspace has a performance
profile that
does not fit well with the default algorithms in MOSES.  A naive port
substituting
template Vertex<>  by Handle could easily slow down MOSES by a factor of
10x or 100x.   Clearly, that's not viable. At least, not to me.

One can get clever in several ways. The most obvious one is to have a
copyin/copyout system to copy Atoms into moses-internal format, run MOSES
for a while, and then copyout back into the atomspace.  There are other,
more
subtle approaches possible.


However as Linas notes, the intersection btw Atomspace and graph-DBs
> is not about MOSES in particular....
>
> A few of us have been thinking about potentially connecting Atomspace
> to Apache Ignite,


Why not Redis or Riak? The issue
https://github.com/opencog/atomspace/issues/1502

describes work that needs to happen prior to a redis/riak/ignite port.


> so that Ignite could serve as a sort of distributed
> backing-store, and the current Atomspace code would serve as the
> "local cache" Atomspace on each machine in the distributed network....
>   Ignite is an interesting beast which is a sort of middleware layer
> between data store and application....   This is not quite the same as
> porting Atomspace logic to a graph DB, but perhaps it's in a similar
> spirit...
>

The actual task to be done depends srtrongy on who is doing it, how
many people are involved, and how clever they are.

I fully expect other opensource groups to reinvent an atomspace-like system
on top of a graph DB. They will start with a clean slate, not reusing any
atomspace
code, and probably reinventing it blisssfully unaware that the Atomspace
exists.

But if you decide to start with the existing code base, then I think
https://github.com/opencog/atomspace/issues/1502
spells out the preliminary groundwork that has to be accomplished, before
you can layer it on top of Ignite (or redis, or whatever).


>
> -- Ben
>
>
>
> On Fri, Feb 2, 2018 at 2:50 AM, Linas Vepstas <linasveps...@gmail.com>
> wrote:
> > Not sure why you say "moses" .. it would be "atomspace" to me.
> >
> > Atomspace has been a kind-of graph database since forever, but there was
> no
> > promotion of it as such on internet forums, and so no one ever heard of
> it,
> > and things like spark/haddoop have taken over 99.99% of the mindshare and
> > actual usage.   What to do in that world?  The atomspace still offers
> some
> > neat ideas and concepts, that these other systems can't yet do ... but
> its
> > only a matter of time.   So I'd say: port the atomspace to these graph db
> > systems.
> >
> > --linas
> >
> > On Thu, Feb 1, 2018 at 10:01 AM, Mark Waser <mark.wa...@gmail.com>
> wrote:
> >>
> >> https://www.oreilly.com/ideas/fishing-for-graphs-in-a-hadoop-data-lake
> >>
> >>
> >>
> >> So . . . . who wants to port MOSES?
> >>
> >> --
> >> 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 opencog+unsubscr...@googlegroups.com.
> >> To post to this group, send email to opencog@googlegroups.com.
> >> 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/00bd01d39b75%
> 24e5a570a0%24b0f051e0%24%40gmail.com.
> >> For more options, visit https://groups.google.com/d/optout.
> >
> >
> >
> >
> > --
> > 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 opencog+unsubscr...@googlegroups.com.
> > To post to this group, send email to opencog@googlegroups.com.
> > 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/CAHrUA36ScSdVJdgBr%3DmhPPz27-
> S-JYOAsLcocTEXgibrWsTpuw%40mail.gmail.com.
> >
> > For more options, visit https://groups.google.com/d/optout.
>
>
>
> --
> Ben Goertzel, PhD
> http://goertzel.org
>
> "In the province of the mind, what one believes to be true is true or
> becomes true, within certain limits to be found experientially and
> experimentally. These limits are further beliefs to be transcended. In
> the mind, there are no limits.... In the province of connected minds,
> what the network believes to be true, either is true or becomes true
> within certain limits to be found experientially and experimentally.
> These limits are further beliefs to be transcended. In the network's
> mind there are no limits." -- John Lilly
>
> --
> 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 opencog+unsubscr...@googlegroups.com.
> To post to this group, send email to opencog@googlegroups.com.
> 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/CACYTDBfgoj_4vb65EtwDrBERkEh-_s8F%2BWMAk%
> 3DcqFHwomFbbEA%40mail.gmail.com.
> For more options, visit https://groups.google.com/d/optout.
>



-- 
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 opencog+unsubscr...@googlegroups.com.
To post to this group, send email to opencog@googlegroups.com.
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/CAHrUA3796euGCfW_c0LENc7NiP7nVDAfSM07JXaHaXj%3DAc2BxA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to