RE: SUO: 21 May 2002 -- Unanswered Questions About SUMO Set Theory
Chris,
There are any number of reasons why a company with a product that is
indeed capable of solving a certain technical problem might not have wide
acceptance. I'm *not* ascribing any of these to Cycorp or any other
company, just trying to show why your argument does not appear sound to me:
1. Hardware isn't powerful enough yet to make the solution high performance
2. Vendor is charging too much
3. Customers aren't happy with Vendor's licensing
4. Vendor doesn't have enough, or cheap enough, or talented enough
application engineers
5. Product was over-hyped
6. Product marketing not effective
7. Product sales team not effective
8. Product documentation is not sufficient
9. Product lacks effective support
10. Any one of the above was true in the recent past, even though it's
good now
There are also semi-technical reasons why a product might "fail" even
though a core idea is a good one:
1. Poor coding resulting in bad system performance
2. Poor coding resulting in reliability problems
3. Bad user interface
..and all of these "X is not good enough" issues might be simply in
relation to other vendors which present some solution which is better along
one or more of these non-technical dimensions. IBM and Oracle do a lot of
systems integration for example, and have extremely strong attributes in
many of these non-technical areas as well as technical ones.
Past failure of a particular system can only be attributed to a particular
implementation at a particular time by a particular company and may not
bear any relation to the future success of that or other companies.
It's also a stretch to claim the Cyc has failed since they in fact have
lots of application success stories.
Adam
At 12:05 AM 5/23/2002 +0200, Chris Partridge wrote:
>Pierluigi,
>
>You wrote:
>"with all respect due to its proponents, this is a very odd argument
>indeed. For one thing, it presupposes what has been long and often shown
>to be an incorrect assumption, that is, that corporations, customers, etc.
>(i) know what they "really need", and (ii) unfailingly recognize "it"
>when they see it. You don't need to be a raving Troskyite (not that
>there is necessarily anything wrong with that) to believe that (i) and
>(ii) have been over and over empirically disconfirmed."
>
>I am not sure that what Bill or I were assuming was "that corporations,
>customers, etc. (i) know what they "really need" " in the sense you are
>talking about. The corporations know they have a problem - they are not
>stupid - and they know they really need a solution. What they do not know
>yet is what that solution is - they know full well they have not yet found
>what they "really need". If the amount of money spent on integration
>projects is any indication - they regard it as a serious problem. As far as
>I am aware it has been "empirically confirmed" that with serious problems of
>this type, corporations are usually prepared to try solutions proposed to
>them that can demonstrate that they work. And the proposition is that if
>something like your employer (Cyc) has been reasonably visible for a while
>and has tried to persuade corporations to use its product (which I
>understand it has) then it cannot have demonstrated to their satisfaction
>that it is the solution. The argument seems to me unimpeachable. BTW this is
>not meant to denigrate Cyc - merely clarify what is not good for.
>
>Is your underlying point that Cyc (or whatever) is the answer to 3)? That
>this is obvious to you and the other initiates. And that for some reason the
>large number of corporations with projects in this area do not 'get it'
>despite the publicity that surrounds Cyc. And despite attempts by Cyc to
>explain to them, and demonstrate, that it will work. This seems to me the
>odd position. It also seems to me to underestimate the acumen of the large
>number of people with considerable experience working in the area.
>
>It is important to get this straight, because if we have the answer to this
>problem then the issue is just a marketing one. If we do not, it becomes a
>methodological one.
>
>Regards,
>Chris
>
>-----Original Message-----
>From: miraglia@cyc.com [mailto:miraglia@cyc.com]
>Sent: 22 May 2002 20:06
>To: Chris Partridge
>Cc: Bill Andersen; Standard-Upper-Ontology (E-mail)
>Subject: Re: SUO: 21 May 2002 -- Unanswered Questions About SUMO Set Theory
>
>
>On Tue, May 21, 2002 at 09:27:52PM +0200, Chris Partridge wrote:
> >
> > Can I support the point underlying Bill's comment that:
> > >> It seems to be recognized by many people who have chosen to join this
> > group
> > >> that SUMO, OpenCyc or any similar effort will not achieve (3).
> >
> > 3) The SUO will play the role of a neutral interchange format whereby
> > owners of existing applications will be able to map existing data
> > elements just once to a common ontology. This provides a degree of
> > interoperability with other applications whose representations
> > conform to SUO. This entails the SUO being able to be mapped to
> > more restricted forms such as XML, database schema, or object
> > oriented schema.
> >
> > As he says "If you want anecdotal evidence then there is anecdotal
>evidence
> > that Cyc has not been successful at this task. If it were, the whole IT
> > industry would be lining up to buy Cyc software and consulting. This
>would
> > be an unfair point if Cyc hadn't been around for so long. Based on the
>fact
> > that SUMO is essentially a Cyc-like architecture, I don't expect any
>better
> > outcome for it." Though I suspect the truth is a little more complicated -
> > the basic point seems to me correct, and important to recognise.
> >
> > Application integration is currently a big headache for enterprises - with
> > millions if not billions spent on it. Enterprises are eager for a
>solution,
> > so if Cyc-type efforts were the solution then there should be some
> > significant evidence that they are working - not just anecdotes.
>Presumably
> > if SUMO is the answer, then corporations would already be using it to
>solve
> > their problems, and asking for more. Note this is not to denigrate the
>work
> > that has gone into SUMO, just the naive assumption that it will resolve
>the
> > kind of problems noted in 3) above.
>
>Chris, Bill et al.,
>
>with all respect due to its proponents, this is a very odd argument
>indeed. For one thing, it presupposes what has been long and often shown
>to be an incorrect assumption, that is, that corporations, customers, etc.
>(i) know what they "really need", and (ii) unfailingly recognize "it"
>when they see it. You don't need to be a raving Troskyite (not that
>there is necessarily anything wrong with that) to believe that (i) and
>(ii) have been over and over empirically disconfirmed.
>
>Moreover, it's obviously possible to conceive of other explanations of
>the same fact: for instance, that "Cyc-type efforts" so far have not
>really focussed on addressing primarily the "desperate needs" mentioned.
>Usually people make the naive assumption that their efforts will solve a
>problem they have identified, not one they have not clearly formulated.
>
>The interesting question can then be raised about whether similar efforts
>could be retooled and exploited to address such needs. Well, can they? I
>for one would be most interested in hearing while either Cyc or SUMO or
>IFF or... _cannot_ in your view be so retooled. This seems like a clear
>question, both theoretical and practical, which many (though not I, sorry
>to report) here seem well equipped to address, although not one that
>benefits greatly from polling imaginary customers, or so I feel.
>
>So let me try to reframe the issue: what is it about an ontology that
>_would_ make it suitable to address these needs? And what is it with
>existing ontologies that makes them unsuitable? I look forward to reading
>your opinions.
>
>regards,
>
>--
>- - - - * * * * * - - - - * * * * * - - - - * * * * * - - - -
>Pierluigi Miraglia Cycorp, Inc.
>Ontological Engineer 3721 Executive Center Dr.
>(512) 514-2988 Austin, TX 78731
Adam Pease
Teknowledge
(650) 424-0500 x571