Re: Enterprise information system

Kingsley,

I admire and respect your enthusiastic, tireless efforts at
evangelization and your fluent command of the relevant technologies in
this area. I mostly agree with your diagnosis and prescriptions, but I
have a growing suspicion that it is not sufficient to meet the real ICT
needs of the enterprise.

On Fri, 2015-02-27 at 18:19 -0500, Kingsley Idehen wrote: 
> On 2/27/15 4:19 PM, Paul Tyson wrote:
> 
> > I don't know that many "linked data" systems improve much on
> > conventional ones. 
> 
> To that answer, I ask: what is the problem? 
> 
> As I see it, the problem boils down to data access, integration, and
> dissemination. This has to happen on time, in the right form, and
> delivered to the relevant entity. 

Enterprises that still have these problems are stubborn, timid,
unimaginative, or in thrall to troglodytic enterprise software vendors.
Getting data from anywhere on the network to a screen in front of a user
is a solved problem, however messy the details might be.

> 
> All the "conventional systems" I am aware of suffer from a common
> flaw, one  that I refer to as data-silo-fication. This problem ranges
> from big iron to tiny computing devices. 
> 
> What are the ramifications of the data-silo-fication? Degradation of
> the following:
> 
> 1. Agility
> 2. Privacy
> 3. Society.
> 

Agreed.

> > We have enough of data and linking.
> How come? The world is full of data silos that hide behind silky
> promises of convenience and/or outright cognitive dissonance. 
> 
> > What we need to
> > link are the artifacts of mental processes, which are not so easily
> > reduced to "data". 
> 
> Yes! And you already achieve that in the so-called "real world" using
> natural language sentences, one of mankind's most powerful inventions
> [1]. 
> 
> We have used language to encode and decode information for eons. 
> 

I think we lose a lot when we characterize language that way. Before
mathematical logic and information science took over, thoughtful people
characterized language as a system of conventional signs representing
mental states or processes.  Of particular importance to philosophers
was the reasoning mind, the mental entities involved in reason, and the
particular structures thereof which seem to correlate to correct
knowledge of the real world. From the time of Aristotle these entities
have been studied as concepts, propositions, and argument (or
demonstration). (I don't mean the poor emaciated use of these terms as
in mathematical logic, but the rich, robust--and complicated--usage they
have in traditional logic from Aristotle on through Maritain and points
in between.)

> > That is the real promise of these technologies, but
> > is not, so far as I am aware, being pursued anywhere in public.
> 
> Well, we have this wonderful thing called the World Wide Web. Its
> basic architecture boils down to:

> 1. Using HTTP URIs as names for entity types (or classes) -- the
> *nature* of entity relationship participants
> 2. Using HTTP URIs as names for predicates (sentence forming
> relations) -- the *nature* of entity relationship types (functional,
> inverse-functional, transitive, symmetrical etc..)
> 3. Using HTTP URIs as names for instances of entity types -- actual
> entity relationship participants.
> 
> In addition to the above, albeit not immediately obvious in HTML
> (which has had <link/> and the <a/> control  in place forever) , the
> aforementioned architecture also included the ability to construct
> sentences where the subject, predicate, and object (optionally) are
> identified by HTTP URIs. 

Yes, I've absorbed all the philosophy of hyperlinking from Vannevar
Bush, Doug Engelbart, Ted Nelson on through Hytime, HTML[2-5], XML &
friends, RDF, and Linked Open Data. I'm here to tell you it's all good,
but *it's not enough*.

> 
> The digital sentences described above can be written to documents
> using a variety of notations (HTML, XML, CSV, TURTLE, JSON, JSON-LD
> etc..), and served from any location on an HTTP network, with name
> interpretation (description lookup or de-reference) baked in. 
> 
> The semantics of the predicates that hold these sentences together are
> both machine and human comprehensible. Thus, anyone can lookup the
> *nature* of an entity relationship type en route to understanding the
> meaning of a given entity relationship.  What more do you need?

Concepts, propositions, and arguments (of traditional logic).

> 
> 
> In my experience  I see a big problem that boils down to understanding
> that "over automation" is bad. Recent fixation with imperative
> programming and applications, in every situation, is utterly broken. 
> 
> The assumption that end-users are dumb, stupid, even lazy is the
> eternal blind spot that afflicts those that simply cannot see, or even
> describe a digital computing realm without fixating on a specific
> programming language, framework, library, data serialization format,
> or some new dogma (e.g., Open Source -- which doesn't guarantee data
> de-silo-fication ).
> 

No doubt.

> 
> In my eyes, what we need is the ability to put language to use, in the
> medium that the Web provides. That simply boils down to systematic use
> of signs [HTTP URIs], syntax [S,P,O or E,A,V term arrangement rules],
> and semantics [meaning of subject, predicate, and object relationship
> roles] for encoding and decoding information [data in some context].
> Basically,  an ability to write the aforementioned digital sentences
> to HTTP network accessible documents, using a variety notations. 

I see this as a Procrustean bed that chops off the better part the human
mind. To be sure, HTTP URIs are a valuable contribution. But they must
be put together in more advanced structures than S,P,O or E,A,V to
represent useful artifacts of the reasoning mind.

> 
> BTW -- if you know of some existing alternative to what I've
> described, that doesn't include a hidden data-silo-fication tax, I am
> all ears :) 
> 
No, I don't know of an alternative. I am just growing more doubtful
about the promise of linked data/semantic technology in its current form
and direction, particularly in the context of enterprise information
systems.

Regards,
--Paul
> 
> 
> 
> 
> > 
> > Note, however, the recent release of Linked Data Platform as a W3C
> > standard (http://www.w3.org/TR/2015/REC-ldp-20150226/). No doubt this
> > will be useful in its own right, but also point the way to future
> > opportunities by what it *doesn't* cover.
> 
> LDP simply addresses read-write issues for those that can't make use
> of SPARQL 1.1 Update, SPARQL Graph Protocol, or WebDAV (which is XML
> specific in regards to metadata). 
> 
> Links:
> 
> [1] http://www.slideshare.net/kidehen/understanding-29894555/55 --
> Natural Language & Data 
> [2] https://www.pinterest.com/pin/389561436488854582/ -- What sites
> between you and your data
> [3] http://kidehen.blogspot.com/2014/07/nanotation.html -- Nanotation
> (inserting digital sentences wherever plain text is allowed)
> [4]
> http://kidehen.blogspot.com/2014/08/linked-local-data-lld-and-linked-open.html -- Linked Local Data vs Linked Open Data .
> 
> -- 
> Regards,
> 
> Kingsley Idehen       
> Founder & CEO 
> OpenLink Software     
> Company Web: http://www.openlinksw.com
> Personal Weblog 1: http://kidehen.blogspot.com
> Personal Weblog 2: http://www.openlinksw.com/blog/~kidehen
> Twitter Profile: https://twitter.com/kidehen
> Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
> LinkedIn Profile: http://www.linkedin.com/in/kidehen
> Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this

Received on Monday, 2 March 2015 23:56:57 UTC