On 1/14/07, Chuck Esterbrook <[EMAIL PROTECTED]> wrote:
* What do you feel the pros and cons are vs. Cyc?
Officially, Cyc is intended as an ontology, not a repository of facts. Although in ResearchCyc they have additional facts. If possible, we may incoporate Cyc's knowledge, perhaps by licensing with them.
I think the main difference is that we make this open to the public.
* How will the system deal with the fact that a bird with broken wings cannot fly? Should the statement be augmented, a new statement introduced, or all statements to be understood as "usually, but not guaranteed"? Or something else?
Yes, the fact "birds can fly" would be marked as *defeasible*. Which is different from tautologies such as modus ponens or exclusion of the middle.
* Would the system explicitly capture probability? "Most personal computers have hard drives."
Yes, that can be done too, eg "50% of marriages end in divorce." But not many facts have probabilities available, especially as a public consensus. If not, they can be marked simply as defeasible.
* Would it support separate domains/modules?
Initially it should be a single common sense body of knowledge. Later, perhaps we'll make special domains.
* Technically speaking, what's the format? Custom text. XML. Maybe via HTTP RPC. Live SQL database. Something else.
Not sure yet. XML seems a good choice. It depends on whether we use FOL or lojban etc.
* What do you feel the most direct applications are for this new repository? (If the same as Cyc, it would suffice to say so. They have a list on their front page at http://www.cyc.com/)
If you ask me, it's for AGI builders =) Anyone can download it for their own purposes, maybe with a fee. YKY ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
