Wednesday, April 30, 2008

Agent theories redux

My sometime correspondent Roy Simpson asked me to comment on the following text from a European research roadmap project. 'What does it mean?' he asks.

"We need agent solutions for distributed, enterprise-wide environments with exacting development requirements. This might be achieved through transition approaches by which existing systems can be upgraded with a successively increased agent presence in a seamless fashion.

"Wrapping legacy systems within autonomous agents situated in a larger multi-agent system is one approach that is being tried, for example, in connecting new and old telecommunications switches together seamlessly, allowing legacy switches to be gradually replaced without major disruption to the overall system."

My answer

Well, where do I start. The term "agent" is not a theoretical term (where is the formal definition?). It's either informal anthropocentric chat from the less reputable side of AI research, or it's whatever the formal system under consideration wants it to mean. Or it's just marketing-speak.

It is a common practice in telecoms, as in computing, to wrap legacy systems with mediation interfaces to talk to more modern systems. The latest incarnation is to provide XML-based web services wrappers around legacy applications. We saw a similar thing with three-tier architectures where the first two tiers 'wrapped' legacy enterprise information systems for the new Internet age (and browser access).

The material you quote is basically gibberish, which could only find a home in a European funding application or as an output from sub-standard and mediocre euro-research.

Saturday, April 26, 2008

Elves and Dwarves

A message has been received in coded neutrino pulses from the stars. A Manhattan-style project has been set up in the strictest secrecy to try to decode it. After months of frustration, a brilliant mathematician is invited to join. This is the plot of Stanislaw Lem’s His Master’s Voice.

The mathematician finds that the government has cast a wide net. The practitioners of hard science have been joined by scholars from all the relevant humanities. And ‘relevant’ has been widely-drawn indeed. The scientists describe themselves as dwarves: the humanities guys they call elves. When the latter put up notices advertising their seminars with the word “science” in the title, the dwarves deface them by appending “-fiction”. And they are not wrong.

Robert Milne, a colleague back in Standard Telecommunications Laboratories in the 1980s, used to distinguish what he called precise thinkers and fuzzy thinkers.

More elf examples.

- People who say that Darwin was mostly right, but has to be amended insofar as his theory doesn’t account for mankind being divinely ordained as the centre of the cosmos.

- People who believe that everyone is exactly the same “really” but that poor and discriminatory environments cause all social problems and it’s only political selfishness that prevents such problems being fixed once and for all.

- And of course the apocryphal “You gotta slow down, Jim. The crew canna take it much longer” from Dr McCoy.

What elf-think has in common is that the arguments sound appealing, are based on desirable human values, and fall to pieces on closer examination. Not that the elves will ever agree.

We have an elf vs. dwarf situation in my present assignment and I was trying to think how it plays out in Myers-Briggs terms*.

I came to the conclusion that dwarf/elf is just a polite synonym for the NT/NF Rational/Idealist distinction (Thinking or Tough-Mindedness vs. Feeling). This explains why there are more women elves and more male dwarves, why politicians who are values-driven are often woefully short of effective policies, and why the truth often hurts.
* The Myers-Briggs community is strongly elven, while the "five-factor model" academics are ferociously dwarvish. This explains some of their mutual animosity.

Friday, April 25, 2008

In the Greens

Now back in Dubai, but in an apartment in The Greens, about 10 minutes walk from the office.

In my two weeks back in the UK, I was able to immerse myself in Maxwell's equations and get my first Open University assignment completed (SMT359 - Electromagnetism). Now that I'm working 12 hour days with the client again, staying on track with course book two is as hard as ever.

I had a quick peek at the end of this particular unit, where it discusses electromagnetic theory and special relativity and introduces the electromagnetic tensor*. As with all good notations, there is a suggestion of a deeper reality lying behind the symmetries of the tensor. Unfortunately, the text stops there!

I'm working on a next-generation network design which conceptually stretches all the way from service-oriented architecture and IMS at the top through an IP/MPLS and fixed/wireless broadband network down to the fibre itself.

It occured to me that as I occupy the network middle ground, Maxwell's equations are defining the way the light transits the fibre (bottom of the stack) while cognitive and psychological sciences are defining the highest level of application functionality (smart application systems and users) at the top of the stack.

The difficulties of combining non-trivial physics with full-time consultancy has made me reflect a little on my plans for next year. I was going to do the quantum mechanics course, but without my full attention it might be smarter to take the less mathematical relativity course and postpone QM a couple of years to when I might be a little less busy.
*As is often the case, the Wikipedia article is too hard.

Wednesday, April 16, 2008

A Sumerian Observation of the Köfels Impact Event

Alan Bond and Mark Hempsell have produced a thorough, scholarly book here which is centred around analysis of the cuneiform tablet K8538 (the "Planisphere") at the British Museum. The tablet seems to be a 700 BC copy of an astronomical clay 'notebook' dating back to the pre-sunrise of 29th June 3,123 BC in Kish (modern Iraq).

Most of the early chapters analyse the Planisphere a sector at a time, decoding the star signs and Sumerian commentary by matching constellations and planets against state-of-the-art programs which can reproduce the night sky from any location and at any time in the last few thousands of years. But it's what happened next which captures the interest. Just before dawn the astronomers see a bright object appearing from behind the clouds, moving faster than anything they have ever seen before. It's big enough to show phases as it traverses the sky from East to West (at this stage it is still exo-atmospheric) before it vanishes into the earth's shadow and then below the Sumerian horizon to the North-West.

Thirty seconds later (p. 71) the object enters the top of the atmosphere over Greece and an ionised shock-front becomes visible as a gigantic arrow head in the pre-dawn darkness. As the object traverses the coast of Albania it is seen as an enormous fireball, larger than the sun.

The asteroid dumps enormous amounts of energy into the atmosphere. As it reaches the southern Alps, it is releasing the equivalent of a one kiloton atomic explosion every metre while the shockwave over-pressure on the ground amounts to several atmospheres.

In its final second of flight, the asteroid gouges a 2 kilometre cut out of the mountain overlooking Austrian Köfels, at the 6 degree gradient of its very shallow trajectory. This glancing blow was sufficient to vaporise the asteroid.

The final 4 kilometre-wide, 850 million tonne fireball hit the mountain at Köfels at 14 km/sec. The energy released was 14,000 Megatons, producing not a crater, but the demolition of the mountain into the gigantic landslide we see today.

The aftermath was an ejecta plume which spread backwards, recapitulating the entry trajectory. This searingly hot mass would have launched into space, achieving a height of 900 km over the Mediterranean, and re-entering over northern Egypt and Syria. The astronomers of Kish saw the plume like a gigantic tree, low on the horizon, a few minutes after their strange star had vanished.

As it re-entered, the plume dumped 500 Megatons of energy into the atmosphere (= four simultaneous detonations of the largest H-bombs ever tested). This would have incinerated anyone beneath it. It also seems likely there were major climate changes post-impact.

Bond and Hempsell tell a thorough story although inevitably there is always a great deal of interpretation with such ancient and fragmentary material. But as they observe, it all seems to hang together.

Sunday, April 13, 2008

15 second recrimination

Niall and Clarissa blagged their way to the top of the Burj Dubai, where they were rewarded with a great view.

Peeking out at the 600 metre drop from a half-finished balcony, Niall toppled off, dragging Clarissa with him.

"Omigod, what an awesome view!"

"Niall! So careless ..!"

He reaches out to embrace her, but she draws back.

"Hey, Clarissa, we're gonna die. No sense entering eternity in a huff."

Turns her face away. "No, I've got to act like I feel. I don't feel like it."

To himself. "View's not so good anymore, either."

"Aw, Clarissa ..."

Friday, April 11, 2008

'The Stuff of Thought' - Steven Pinker

Just finished Steven Pinker's mammoth "The Stuff of Thought" which was a Christmas present. At 512 pages, I don't have the energy to review it properly - there are plenty of good reviews on the Amazon site - follow the link.

Let me just say that you probably have to have an interest in linguistics to read it - as general science or even evolutionary psychology, it lacks that page-turning magic. I liked the many occasions when he observes that authority figures use self-deprecatory language to increase a sense of communalism with their audience - and then the many other occasions when self-deprecatory stories about Steven Pinker are aired. He knows that you know that he knows why he is doing this.

He calls his position "conceptual semantics" and it seems hard for reasonable people to disagree with him. I guess my bottom line view would be that the book as a whole seems eclectic. In the end, there is going to be a coherent story founded on sociobiology and rooted in evolution: we're not there yet.

I'm still in the UK writing up the work I was doing with the client in Dubai. This seems likely to take up most of April. Having contracted a rather nasty respiratory bug out in Dubai, I can report that I'm still a bit deaf and my voice is not back to normal. However, it's a lot better than it was ... I was capable of work again from Tuesday of this week.

Thursday, April 03, 2008

The Köfels event

This is a truly amazing story from here.

A cuneiform clay tablet that has puzzled scholars for over 150 years has been translated for the first time. The tablet is now known to be a contemporary Sumerian observation of an asteroid impact at Köfels, Austria and is published in a new book, 'A Sumerian Observation of the Köfels' Impact Event.'

The giant landslide centred at Köfels in Austria is 500m thick and five kilometres in diameter and has long been a mystery since geologists first looked at it in the 19th century. The conclusion drawn by research in the middle 20th century was that it must be due to a very large meteor impact because of the evidence of crushing pressures and explosions. But this view lost favour as a much better understanding of impact sites developed in the late 20th century.

In the case of Köfels there is no crater, so to modern eyes it does not look as an impact site should look. However, the evidence that puzzled the earlier researchers remains unexplained by the view that it is just another landslide.

This new research by Alan Bond, Managing Director of Reaction Engines Ltd and Mark Hempsell, Senior Lecturer in Astronautics at Bristol University, brings the impact theory back into play. It centres on another 19th century mystery, a Cuneiform tablet in the British Museum collection No K8538 (known as "the Planisphere").

It was found by Henry Layard in the remains of the library in the Royal Place at Nineveh, and was made by an Assyrian scribe around 700 BC. It is an astronomical work as it has drawings of constellations on it and the text has known constellation names. It has attracted a lot of attention but in over a hundred years nobody has come up with a convincing explanation as to what it is.

With modern computer programmes that can simulate trajectories and reconstruct the night sky thousands of years ago the researchers have established what the Planisphere tablet refers to. It is a copy of the night notebook of a Sumerian astronomer as he records the events in the sky before dawn on the 29 June 3123 BC (Julian calendar).

Half the tablet records planet positions and cloud cover, the same as any other night, but the other half of the tablet records an object large enough for its shape to be noted even though it is still in space. The astronomers made an accurate note of its trajectory relative to the stars, which to an error better than one degree is consistent with an impact at Köfels.

The observation suggests the asteroid is over a kilometre in diameter and the original orbit about the Sun was an Aten type, a class of asteroid that orbit close to the earth, that is resonant with the Earth's orbit. This trajectory explains why there is no crater at Köfels. The in coming angle was very low (six degrees) and means the asteroid clipped a mountain called Gamskogel above the town of Längenfeld, 11 kilometres from Köfels, and this caused the asteroid to explode before it reached its final impact point. As it travelled down the valley it became a fireball, around five kilometres in diameter (the size of the landslide). When it hit Köfels it created enormous pressures that pulverised the rock and caused the landslide but because it was no longer a solid object it did not create a classic impact crater.

Mark Hempsell, discussing the Köfels event, said: "Another conclusion can be made from the trajectory. The back plume from the explosion (the mushroom cloud) would be bent over the Mediterranean Sea re-entering the atmosphere over the Levant, Sinai, and Northern Egypt.
"The ground heating though very short would be enough to ignite any flammable material - including human hair and clothes. It is probable more people died under the plume than in the Alps due to the impact blast."


I am now back in the UK with another few weeks to go on my Dubai contract. I am also still in recovery from an extremely unpleasant cold during which I lost my voice. My presentation to senior client people two days ago remains particularly vivid in my mind. Flying with a heavy cold is also particularly not recommended.