Re: Justifying decentralization: Epistemology in the Cloud

On 25 Aug 2018, at 22:41, Kingsley Idehen <kidehen@openlinksw.com<mailto:kidehen@openlinksw.com>> wrote:

On 8/25/18 8:25 AM, Story H.J. wrote:
My paper for Decentralizing the Semantic Web at ISWC 2018 has been accepted. Here is the abstract.

Abstract: The web is an open platform that allows anyone to publish anything, and so raises anew many epistemological questions: how can one distinguish what is true, what is fake or what is fictional on the web? Indeed how can one know anything at all? We start from an analysis of knowledge that makes space for radical skepticism and which allows us to locate the essential problem with the current web application architecture. This allows us to propose a set of criteria that explicate and justify the decentralised architecture of the internet and the web, and the need for that to be extended to the data and application layer. The proposed architecture is socio-technical, recognising the roles of individuals, institutions and nations in our epistemic makeup. We illustrate this by proposing an architecture of trust that ties these institutions into browsers in a decentralised and open way, allowing them to make the web a more trustworthy space. As a side effect we gain the tools to make some  serious inroads in helping combat Phishing.

The paper is up here with links to further blog posts that elaborate on different aspects of it.
https://medium.com/@bblfish/epistemology-in-the-cloud-472fad4c8282


One of the articles linked to there "Stopping (https) phishing", makes the point that everyone
has moved to https even the Phishers. So we should too. Here is an updated image for the WebID
spec.


<WebID-overview.png>


Hi Henry,

The document (identified by https://www.w3.org/People/Berners-Lee/card) describes an entity (identified by https://www.w3.org/People/Berners-Lee/card#i). Your "describes" relation currently has a photo as its object.

Well actually it is pointing into the photo at Tim, just as the sense meaning relation is pointing into the
graph.


Your illustration also indicates (from a general perspective) that https://www.w3.org/People/Berners-Lee/card#i denotes a photo.

It points into the photo. Now the photo could be a lot better, higher resolution. One could remove the background...

What is needed is to extend the spec by bringing in some technical concepts that I have gotten by developing stuff
with Alexandre Bertails in Scala, and that I am want to define carefully in category theoretic terms for my thesis. I will give in my second year report end of September, and it should be reviewed thereafter. But I may try to blog it when I am writing up the thesis to get some feedback, and that could then be used to give some mathematical substance
to the WebID spec.

So there is the need to define the following concepts in addition to the well known RDF Graph.
We need a PointedGraph, which is just a pair Node × RdfGraph. There is a simple algebra that allows one to move
through a PointedGraph by following arcs. Actually it may be a coalgebra because it allows one to observe the graph.
Lets Abbreviate PointedGraph to PG, have 𝒫 be the PowerSet functor, and ^ the power operator where 2^3 would be 8,
but with types A^B it means the functions from B to A.

  explore: PG → 𝒫(PG)^URI

explore is the function that given a PG returns a function from URIs to sets of PGs.

Now we can extend that to NamedPointedGraphs which are pairs of the form  URI × PG, and so isomorphic
to the Triple URI × Node × Graph. This allows us to distinguish where we got a graph from, since it makes
a big difference where we got it from. Here we also have a coalgebra

  explore: NPG → 𝒫(NPG)^URI

But here it is possible to jump out of a graph into another graph on another server...
That function is not relevant to the argument here though, so I won't go into it.

What is important is sense is the following function

sense: URL → NPG

It is a function that if it resolves gives us a NPG. But it does not give us the same NPG at every
request - since the web changes. So it really is a function of the type

sense: Web →   (Web×NPG)^URL

where the Web is something stateful that changes over time. This again is coalgebraic because we get
a new state of the web after a call into it. In OO notation this is written as

val timbl = web.sense(url)

timbl here is the pointer in the named graph that we see in the picture below. That is why there is a point there
on Tim.

Now describes is a relation from a PG to an object in the world. This is related to semantics. It has to be
a relation from a PG to the object in the world, because the graph itself maps to a pattern in the world: one in which
all the relations have the same shape as that in the graph. But it does not pick out any object in particular. But the
pointed graph can pick out an object in particular, in the graph as in the world...

Anyway I think all of that needs to be added to the spec at some point, but I think it would be wise if I write it out carefully and present it for review to some mathematicians/logicians first, before putting it in the spec.

Still we should update the spec at some point and a better picture of Tim may be useful. Perhaps the functorial nature
of semantics could be made clearer by having a picture where the relations between objects in the picture mirror those
in the graph... We would need some definite description of Tim but by relating him to some objects that mirror those in the graph...

[cid:03AD47F0-EB56-47EE-9621-3A7C2DC34293@lan]

I know you understand all of this.

The main challenge here is trying to produce an illustration includes the "minds eye" mark for the entity identified by the HTTP URI https://www.w3.org/People/Berners-Lee/card#i, bearing in mind powerful (but generally confusing) denotation->connotation quality of an HTTP URI courtesy of the "#" indexical .

Here's what you are trying to illustrate (plus some additions by me so that you can lookup a variant of your illustration) .

## Turtle Start ##

@prefix wdrs: <http://www.w3.org/2007/05/powder-s#><http://www.w3.org/2007/05/powder-s#> .
@prefix schema: <http://schema.org/><http://schema.org/> .
@prefix foaf:  <http://xmlns.com/foaf/0.1/><http://xmlns.com/foaf/0.1/> .
@prefix timbl-card: <https://www.w3.org/People/Berners-Lee/card><https://www.w3.org/People/Berners-Lee/card> .
@prefix timbl: <https://www.w3.org/People/Berners-Lee/card#i><https://www.w3.org/People/Berners-Lee/card#i> .
@prefix kidehen-preso: <https://www.slideshare.net/kidehen/how-virtuoso-enables-attributed-based-access-controls/8><https://www.slideshare.net/kidehen/how-virtuoso-enables-attributed-based-access-controls/8> .
@prefix depiction: <https://upload.wikimedia.org/wikipedia/commons/8/83/Tim_Berners-Lee-Knight-crop.><https://upload.wikimedia.org/wikipedia/commons/8/83/Tim_Berners-Lee-Knight-crop.> .

timbl: wdrs:describedby timbl-card: , kidehen-preso:  .

timbl: foaf:depiction depiction:jpg .

timbl-card: a foaf:Document, schema:CreativeWork  ;
         foaf:primaryTopic timbl: ;
         schema:mainEntityOfPage timbl: .

kidehen-preso: a foaf:Document, schema:CreativeWork  ;
         foaf:primaryTopic timbl: ;
         schema:mainEntityOfPage timbl: .


## Turtle End ##


BTW -- We have a browser extension called the Structured Data Sniffer [1] that does assist with the visualization of RDF relations e.g., the kind of illustration you are pursuing here.

Links:

[1] https://osds.openlinksw.com<https://osds.openlinksw.com/> -- Home Page with links to Chrome, Firefox, Opera etc., extensions
[2] https://chrome.google.com/webstore/detail/openlink-structured-data/egdaiaihbdoiibopledjahjaihbmjhdj?hl=en -- Chrome Extensions
[3] https://addons.mozilla.org/en-US/firefox/addon/openlink-structured-data-sniff/

[4] https://www.pinterest.com/pin/389561436515442506/ -- Pintrest Pin demonstrating visualiztion of the RDF-Turtle example above.


--
Regards,

Kingsley Idehen
Founder & CEO
OpenLink Software   (Home Page: http://www.openlinksw.com<http://www.openlinksw.com/>)

Weblogs (Blogs):
Legacy Blog: http://www.openlinksw.com/blog/~kidehen/

Blogspot Blog: http://kidehen.blogspot.com<http://kidehen.blogspot.com/>
Medium Blog: https://medium.com/@kidehen


Profile Pages:
Pinterest: https://www.pinterest.com/kidehen/

Quora: https://www.quora.com/profile/Kingsley-Uyi-Idehen

Twitter: https://twitter.com/kidehen

Google+: https://plus.google.com/+KingsleyIdehen/about

LinkedIn: http://www.linkedin.com/in/kidehen


Web Identities (WebID):
Personal: http://kingsley.idehen.net/public_home/kidehen/profile.ttl#i

        : http://id.myopenlink.net/DAV/home/KingsleyUyiIdehen/Public/kingsley.ttl#this