Code

Programming Languages: Lisp, Python, Java

License: Apache License V2.0

Repositories

browse code, statistics cvs -d:pserver:anonymous@opencyc.cvs.sourceforge.net:/cvsroot/opencyc login

cvs -z3 -d:pserver:anonymous@opencyc.cvs.sourceforge.net:/cvsroot/opencyc co -P modulename

Show:

What's happening?

  • Followup: RE: processing the Queue

    parse-a-question-completely will not be available for ResearchCyc for now. We will be releasing documentation for our OpenCyc Java APIs soon. Please stay tuned!

    2013-05-16 12:04:39 PDT by vijayrajcyc

  • Followup: RE: processing the Queue

    I should probably clarify that these questions are about researchCyc, not OpenCyc.

    2013-05-14 03:59:18 PDT by https://www.google.com/accounts

  • Followup: RE: processing the Queue

    Actually, those were the answer to question 2. The answer to question 3 is "generate-text-w/sentential-force." All of these seem to be working except parse-a-sentence-completely and parse-a-question-completely, which always seem to return NIL, even on the example sentences: (parse-a-question-completely "What countries border on Turkmenistan?" #$RKFParsingMt...

    2013-05-14 03:56:28 PDT by https://www.google.com/accounts

  • Followup: RE: processing the Queue

    I just found the answer to question 3 above: ps-get-cycls-for-phrase ps-get-cycls-for-np and parse-a-sentence-completely.

    2013-05-13 09:59:39 PDT by https://www.google.com/accounts

  • Followup: RE: processing the Queue

    Interesting, I hadn't noticed that difference (between NIL and (NIL)). If you don't mind a further question, what Cyc-L commands should I use to obtain 1. the steps of a proof (used to be "fi-prove" I think) 2. the parse of an English phrase (used to be "parser" I think) 3. The natural language version of a statement (the kind you can get in the KB browser window)

    2013-05-13 09:20:24 PDT by https://www.google.com/accounts

  • Followup: RE: processing the Queue

    FI-ASK bottoms out to NEW-CYC-QUERY. "new-cyc-query" should be used to return results. The fully bound query will return true or could not prove. For true "(NIL)" is the data structure returned by new-cyc-query. When it can't prove, we return "NIL". Note the only difference is NIL vs singleton NIL, i.e. NIL within parenthesis.

    2013-05-10 05:27:55 PDT by vijayrajcyc

  • Followup: RE: processing the Queue

    In this case I could get the answer in various other ways using the Java API, I just am trying to understand why CycL behaves the way it does. For example, I would like to be able to get back a chain of reasoning, but the only way I have found to do that is by using PROVE in CycL, which behaves the same way (returns :QUEUED.)

    2013-05-08 10:36:32 PDT by https://www.google.com/accounts

  • processing the Queue

    I tried to reproduce the sample API session from here: http://www.cyc.com/documentation/sample-api-session It works correctly up to a point, but on this line: (fi-ask '(#$isa #$Fido #$Mammal) #$BaseKB nil 2) it returns :QUEUED I see the command show up in the Master Transcript a few seconds later. How do I tell the server to go ahead and process the queue? I set the following...

    2013-05-08 10:02:42 PDT by https://www.google.com/accounts

  • Followup: RE: thereExistAtLeast

    Let's call this a bug. Will be fixed in a subsequent release.

    2013-04-26 10:16:07 PDT by larrycyc

  • thereExistAtLeast

    The example given is (thereExistAtLeast 7 ?X (isa ?X Sea)) yet when I try with an obviously wrong number: (thereExistAtLeast 700000000000 ?X (isa ?X Sea)) It says the Query is True because Full Justification : (isa ArcticOcean Ocean) in WorldNaturalGeographyVocabularyMt (genls Ocean Sea) in UniversalVocabularyMt Is there a bug here?.

    2013-04-26 08:12:35 PDT by https://www.google.com/accounts

Our Numbers