Friday, March 07, 2008

Elements of interaction

I have in the past identified the developments of Web Science in a previous posting. Recently, I came across the following position paper by R.Milner in the workshop about semantic web directed by Hendler in 2005. There many things that I want to comment about but for the time being, I pick as the most important ones its reference to the term Computing-as-interaction and the statement it makes about describing events and activities taking place over the web.
The paper introduces a three-layered scheme, originating from H.Simon's Theories of the Artificial, that can be used as a reference model to successfully describe activities such as "web-clicks".
I dare say, I started in late 80s something similar, equipped with much weaker armoury, the idea to understand the Net in terms of being an evolving computing entity. My prime concern was to locate some basic concepts at first and later to map upon them the emergence of Net advances. So, I consider the position paper above as a kind of validation of my approach .

Now it is for the benefit of fixing the semantic web requirements, back then it was fixing my comprehension needs, modulated to think "holistically" about computing artefacts in general, hence I did the same about the Net itself. This way of thinking owes alot to the great works of D. Turner that I became familiar with in late 70s and later due to a very similar exploration of computing paradigms, pursued by V. Turchin (aka Russian-LISP inventor). I discovered it while on functional programming in early 80s. I tookit up more seriously as I entered Networking in late 80s. I think it was Steve Kille (X.500 isode) mentioning ASN.1(+) that made me apply Turchin logic (evolutionary functional programming) to the problem of developing an undertanding
about the Network.
The quantum of automation (QofA) owes a lot to combinators (#) each S,K,I definining the simplest of automation steps. Initially QofA was my way to understand why P.Maritz's work with S-Algol + 8086 board hit the road rather than the work of G.Steele's LISP machine (aka Scheme) in early 80s. I had worked with the latter and expected it to flourish. When PC/DOS made the scene everywhere its success raised an issue for me. (Also for many Computer Scientists that did not pay attention to the micro wave of technology - please mind the blank character). Now we know it is the combination of Moore (Intel hardware) and Gates (MS strategy) that answers the problem. The appropriate QofA goes forward first. Note, also that the inappropriate one did hit the road eventually as Java VM (recarnating Landin's definitional interpreters) and Steele became the first editor of the manual by SUN. So it's a small world definately. So, it is a bit messy how I came to think about QofA as the general pattern of Net advances.

Computing-as-interaction was Licklider's symbiosis, Engelbart's augmentation in mid 60s. My own first exposure to it, was by Turchin's meta-mechanical process (REFAL machine + Users). When I communicated to Turner around '96 about internet and functional programming he made reference to Milner's work (93 Turing award) I think since he said that parallelism needs to be a 1st-class citizen at programming language level not just an implementation technique. Interesting that Milner insists on Elements of interection in 2005.

All the above look very similar (*) to Internet Systematics, as a pattern of concepts to approach the Net. This gives me the idea to review the so called "layered schemes" like OSI, Bakens, End2End and finally analyse my own "Layers of evolution scheme" as a model of events and developments taking place not just over the web, as Milner suggests but on a pre-web as well as a post-web scale.

So, I am very happy to describe internet systematics as a foundational theory about the evolution of Elements of interaction. Of course I am not in the same business as Milner, I am not equipped for such a mathematical task.

The folowing slide show some diagrams that exhibit the thinking pattern associated with internet systematics


(#) combinators is a concept equivalent to Lambda Calculus, a strong point of influence upon R.Milner's Pi Calculus.

(+) ASN.1 abstract syntac notation, a universal language to express all data structures used by communication protocol technology. Turchin's REFAL, had a similar universal role in expressing all kinds processes. This was a signal for me, a match maker to entagle Turchin's ontology for describing Net events, a kind of Net Science .

(*) ok, similary means some common things such as the influence of Lambda Calculus, the goal of locating basic concepts to model to model upon Net events and developments.

0 Comments:

Post a Comment

Subscribe to Post Comments [Atom]

<< Home