SUO: RE: OpenCyc problems
Thanks for the reply, Pat.
I think I remember you bring up the function issue at VerticalNet.
How did Cycorp respond to your concerns? I'd like to see their
responses.
Please see below for some questions and comments.
Cheers!
-Eric
> -----Original Message-----
> From: Patrick Cassidy [mailto:pcassidy@bellatlantic.net]
> Sent: Monday, June 23, 2003 11:36 AM
> Cc: standard-upper-ontology@ieee.org
> Subject: SUO: OpenCyc problems
>
>
> In response to Eric's suggestion that problems with OpenCyc be
> described explicitly:
>
> Eric Peterson wrote:
> >
> > I could have easily missed some, but the only concrete OpenCyc
> > criticisms that I have seen surface on this group, I had to
> > press for. And they turned out not to be valid.
>
> . . . I will provide one minor example.
> There are others. My main concern with OpenCyc is in
> what it leaves out, specifically Processes which are a
> subclass of functions,
[ELP] The OpenCyc IDE doesn't supply constant grepping out of the box
like full-Cyc. And I couldn't find "Process" in an initial substring of
any constant. I need a text version like yours to search, but of the
current version.
To what specific process-related constant are you referring?
> and substances which are more abstract
> than pieces of a particular substance.
[ELP] I'm not sure that I understand this second one. Can you point me
to the offending axioms?
>
> Today's problem is in the definition of "SheetOfSomeStuff"
> and its subclass "PortalCovering". The latter is two
> levels down the class hierarchy from the former.
[ELP] Pat, in OpenCyc 0.7.0 12-15-02 (which is the latest version) SOSS
is not any sort of generalization of PC.
Perhaps they responded to your earlier bug report. Thanks for being
part of the solution. Concrete specific verifiable criticisms are
certainly precious metal in a world of dross.
> The definitions:
>
>
> #$SheetOfSomeStuff:
> OPENCYC 1: MAY 23, 2002
> A specialization of #$PartiallyTangible. Each instance of
> #$SheetOfSomeStuff is a substance which (in its `intrinsic' shape) is
> continuous and, when flattened, uncrumpled, unfurled, etc., has one
> dimension which is significantly smaller than the other two
> dimensions. The requirement of being a continuous sheet does not
> preclude some type of penetrability, as in instances of #$Screen-Mesh,
> #$Netting, or #$LaceCloth. Instances of #$SheetOfSomeStuff may be in
> solid, liquid, or gaseous form. An instance of #$Puddle on a
> relatively flat surface (but not in a deep pothole) or an instance of
> #$AltoStratusCloud would be members of this collection. Note that
> being an instance of this collection implies nothing about the current
> configuration of the instance; in particular, instances of
> #$SheetOfSomeStuff need not be lying flat. For example, aluminum foil
> on a roll or a bedspread crumpled up on a bed are instances of
> #$SheetOfSomeStuff. Cf. #$SheetShaped, #$TwoDimensional.
>
> #$PortalCovering
> OPENCYC 1: MAY 23, 2002
> A specialization of #$SolidTangibleThing. Each instance of
> #$PortalCovering is a solid tangible object that covers a portal.
> Instances of #$PortalCovering may be either flexible or rigid, and
> they may be either fixed in place or removable. Examples include
> windows (in buildings or automobiles), doors, shutters, screens,
> draperies, window blinds, convertible tops, lids, and corks.
>
>
> . . . and #$Portal itself:
> OPENCYC 1: MAY 23, 2002
> The collection of all openings, as in a surface, through a tube, etc.,
> with or without a covering. This includes doors, mouths,
> doughnut-holes, etc.
>
> Notice that the #$PortalCovering is intended to include corks
> (coverings for bottle portals), But a cork doesn't fit the
> criterion for SheetOfSomeStuff (which is a very useful concept).
> So it shouldn't be a subclass. #$Cork itself doesn't appear
> in OpenCyc (at least the version I looked at six months ago).
>
> I suspect that the concept of "Portal" is just too overbroad
> in OpenCyc and could usefully be subdivided into portals in
> rooms in a structure (which are planar) and other kinds of
> portals, like those in containers (corks, stoppers in a flask,
> spigots on the end of a pipe, etc.) Then #$PortalCovering would
> also be divided. Perhaps Cyclists find the very broad
> #$PortalCovering to be useful, but I can't think of a situation
> where I would want to cover all those very different cases
> at the same time. An example of where the general case is
> actually useful would help.
>
> Pat
>
> =============================================
> Patrick Cassidy
>
> MICRA, Inc. || (908) 561-3416
> 735 Belvidere Ave. || (908) 668-5252 (if no answer)
> Plainfield, NJ 07062-2054 || (908) 668-5904 (fax)
>
> internet: cassidy@micra.com
> =============================================