Re: Blank Nodes Re: Toward easier RDF: a proposal

> On 23 Nov 2018, at 14:33, Thomas Passin <tpassin@tompassin.net> wrote:
> 
> On 11/23/2018 8:21 AM, Jürgen Jakobitsch wrote:
>> in general: making something "simplier" (whatever that means) because someone doesn't understand it
>> is the highway to the dark side. no good thing will ever come out of this.
> 
> Making something easier to *use* is, however, a useful thing to do.

That is usually dependent on tooling, rather than theory.

> 
> > keep in mind kant's categorical imperative and ask yourself where does
> > this finally lead to.
> 
> And how many of middle-of-the-road programmers are able to read and understand Kant?  And that's the point, I think.

It often takes quite a high level of abstraction to completely describe something simple.
 Here are a few examples:

- Bertrand Russell's huge Principe Mathematica set out to prove that 2+2=4. We all know how to count, but the relation framework built on set theory required a lot of work -  which was later much simplified.
- One simplification was of numbers in terms of Monoids, a category with one object and infinite numbers of arrows. This ties numbers in with lists in the category of Monoids. A concept that most people have not heard about, though more and more programmers are learning Category Theory.
- There are some very neat descriptions of OO programming in terms of coalgebras. See for example "Coalgebras and Monads in the Semantics of Java" 
   https://www.sciencedirect.com/science/article/pii/S0304397502003663
  That did not stop me learning Java before I knew any of this. One can learn by doing.

Developers learnt a lot more complex things that RDF. 

I think before trying to cure a problem one has to work out what the cause is. RDF itself is extremely simple.
I have taught it to developers quite quickly, and would have been faster if I had libraries that were correctly
types as we now have with banana-rdf.  

The problem was more of devs understanding what the problem that is being solved is. This is the problem of creating hyper-apps, that extend the web into applications that can follow links across institutions. If that is not made clear, the reasons for the use of URIs won't make sense. The SoLiD Project is starting to build such apps. As those become good enough the choices made in RDF will make complete sense. To learn it helps to build, but to build it helps to see a few buildings around and understand what they are for.

Henry

Received on Friday, 23 November 2018 13:55:59 UTC