> Hi. I have a few questions about the Cyclify project announced here:
> http://opencyc.org/
>
> 1. Will the opencyc mailing lists (like this one) be the main discussio=
n forum
> for Cyclify?
That's not currently the plan. People may still simply want to do
development with OpenCyc, so we don't want to intrude on these lists.
Cyclify has a more specific mission, as discussed in the announcement
presentation that was turned into a podcast.
>
> 2. If I find a possible bug in the ontology, will I send it to one of t=
hese email
> lists (or the opencyc bug tracker)?
> http://sourceforge.net/tracker/?atid=3D389632&group_id=3D27274&func=3Dbr=
owse
>
No. We're going to set up more specific workflows within the Cyclify projec=
t.
> 3. Cyclify is compared to the Wikipedia project. Already on the Cyc Fo=
undation page
> http://www.cycfoundation.org/ there is a reference to the concept of "o=
ntology" that
> gives a link to the Wikipedia article. This underscores that the Wikipe=
dia article URL
> is the de facto standard to referring to a concept on the web (not #$Ont=
ology).
Ouch. Actually, it was an expedient way to give neophytes an idea of
what ontology is. If all of the OWL community adopted Wikipedia's
article title as the canonical term, that might be more like a de
facto standard. I actually don't think there is a de facto standard
for this yet.
> What are the plans in Cyclify to work with the fact that the Cyc-L names
> won't be the de facto universal identifiers when a web user wants to lin=
k to a concept?
> Will Cyc adopt a URL system? And map it explicitly to other identifiers=
like Wikipedia?
>
Basically, yes. There will be a URI for every Cyc term. And there will
be mappings. The details will be worked out by Cyc Foundation members
and then shared more formally.
> 4. Wikipedia is already doing ontology stuff. Their Category system al=
ready does a lot
> of what #$isa and #$genls does. And they are planning typed predicate t=
ags. See
> http://wiki.ontoworld.org/wiki/Main_Page .
Yes, a lot of communities are doing ontology stuff. That's our point.
> What is the vision of coordination as both WCyc and Wikipedia (and other=
s) vie to have > web users contribute to the respective (huge) ontologies? =
Since the Cyc language and
Join the Cyc Foundation and help build the vision!
> query engine will always be much more rich than what Wikipedia will attem=
pt, is the
> idea to just have Cyc "scrape" the ontology of Wikipedia? Could there ev=
er be enough
As a first step, we want to extend the breadth of Cyc with concepts
pointed to by Wikipedia article titles. This will at the same time
provide a more useful ontological structure to Wikipedia than they
could develop from scratch.
> friendship between the projects so that a Wikipedia page has a built-in l=
ink to the Cyc
> engine to do serious natural language queries (which Wikipedia will not l=
ikely attempt)?
I think a good friendship will develop, and we're certainly going to
work to make that happen. And yes, I think serious natural language
queries of Wikipedia are a realistic vision of the future, although I
wouldn't expect it in the next two years -- at least not against
article content.
I met with Jimmy Wales a few weeks ago, and he said he is a supporter
of Cyc and would like to see it grow the way Wikipedia did. He said he
sees the projects as quite similar.
>
> Thanks,
> - Jeff
Jeff (and others), I would invite you to continue this discussion
within the Cyc Foundation, which you are all welcome to join. Just
send an email to johndcyc@... requesting membership. It would be
helpful if you could include a summary of your skills, or whatever
version of a resume you have lying around, so that we could see how
best to make use of your skills.
Thanks,
#$JohnD
|