Re: Misconceptions about what knowledge representation truly is

Dave,

Thanks for the PKN reference — I agree that “mathematically 
representable” in the narrow sense is too restrictive.

What I’m really after is knowledge that is structured enough (whether in 
axioms, PKN‑style statements, or other KR notations) that a system can 
reason over it with explicit notions of domain and adequacy, rather than 
treating everything as an undifferentiated vector soup. K3D is designed 
to host both: fully axiomatized bits and heuristic plausibility layers 
like PKN, as different fields over the same Houses and Galaxies.

Best,
Daniel

On 11/18/25 8:01 AM, Dave Raggett wrote:
> I am not convinced that computable knowledge needs to be explicitly 
> mathematically representable. For example, the Plausible Knowledge 
> Notation [1] uses a variety of statement forms plus a small number of 
> heuristics for computing plausibility in lieu of detailed statistical 
> models. This is in respect to imperfect knowledge that is uncertain, 
> imprecise, context sensitive, incomplete and subject to change. This 
> replaces mathematical proof by rational argument.
>
> [1] https://w3c.github.io/cogai/pkn.html
>
>> On 14 Nov 2025, at 20:56, Daniel Campos Ramos 
>> <danielcamposramos.68@gmail.com> wrote:
>>
>> Milton, Dave, all,
>>
>> I resonate strongly with the idea that a central task for this CG is 
>> to narrow down what we mean by “knowledge that is mathematically 
>> representable in machine‑readable format,”
>>
> Dave Raggett <dsr@w3.org>
>
>
>