Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: SUO: Let's focus




At 20:42 2002-04-08, jim.s3@juno.com wrote:

>Bill and SUO Participants,
>
>         ...
>
>         OpenCyc will probably be proposed as a starter document within a 
> few weeks. I'm surprised at the lack of discussions on this.

Jim,

I'd have to say that OpenCyc isn't open. Unless its theorem prover / 
inference engine code is released, it is fundamentally impossible to know 
what the system really is or does. I don't believe reverse engineering it 
is a suitable way to make this determination.

How can it become an IEEE standard (starter, draft, or otherwise) when the 
most important element in attaching semantics to expressions in CycL is hidden?


 From <http://www.opencyc.org/OpenCyc_org/faq/faq1_html>:

"Q: Will the inference engine be open source? Why (not)?

"A: Effectively yes, but not quite 100% at this time. Cycorp intends to 
sell premium products and services using its inference engine, which has 
been designed to work with the Cyc Knowledge Base in an optimal fashion. 
Cycorp provides a reference Cyc Server executable for Intel-based Linux and 
for Windows 2000 along with the release. This includes a basic 
implementation of its inference engine, but no source code. This coupled 
with the Cyc API is almost as powerful as open source code."


The most charitable thing I can say about this, especially the last 
sentence, is that it betrays a deep, fundamental misunderstanding of what 
open source is.


OpenCyc is not open source software. It's hard to see how they got 
resources from SourceForge with a huge gap like this in the source code 
they release.


Randall Schulz
Mountain View, CA USA


>...
>
>Jim Schoening