On Tue, Feb 6, 2018 at 9:47 PM, 'Nil Geisweiller' via opencog < opencog@googlegroups.com> wrote:
> On 02/03/2018 06:29 AM, Ben Goertzel wrote: > >> Nil, what are your thoughts on >> >> https://github.com/opencog/atomspace/issues/1502 >> >> ? >> > > I seem to agree, however I don't think SetLink should be banished, instead > it should probably be parsimoniously used, when for instance the user/agent > wants to represent an immutable concept, probably a small one. > Yes. > > Although, I've been thinking why not replace, at least underneath, > ListLink by ConsLinks (and perhaps SetLink by SetConsLinks)? This would > solve the arity limit on the persistent storage side. For instance > > List A B C > > would be represented by > > Cons A > Cons B > Cons C > List > > and > > Set A B C > > by > > SetCons A ... > > not sure how to deal with the canonical ordering, I suppose the user/agent > would enter a tree of SetCons in any order, and the AtomSpace may > automatically re-order it. > OK, this is actually a major unsolved problem for the atomspace. We need the ability to create ordered lists of things, and the ability to edit those ordered lists. Currently, this is extremely difficult verging on impossible. And I have no answer for it, either. The problem with Cons A Cons B Cons C List is that, if I want do remove C from the sequence, I'd have to first: delete Cons A then delete Cons B Then delete Cons C then recreate Cons B (without the C under it, this time) and then recreate Cons A. Ugh. Lots of work, and it's begging for a race condition. Worse, if Cons A has a non-empty incoming set, then it cannot be deleted without clearing out the incoming set first. Which might be impossible. In essence its quick-setting concrete, and you need a jackhammer to "fix it". Huh. Can I coin a new saying? "fix it with a jackhammer" --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 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/CAHrUA350AbvDu8-cDe%3DTaCaA5bcCRjKk9Ev2CnbdmQ-G4yhTiA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
