I have recently seen a show with Richard Dawkins, the famous proponent
of evolution and atheism, as the main guest. In the course of the show
Dawkins debated other guests and members of the audience regarding
issues such as the existence of God. What struck me most was what a
poor debater Dawkins is. He constantly played to his supporters and
made no credible effort to convince his opponents they are wrong. Of
course, he told his opponents they are wrong, and he made strong case
to his supporters that his opponents are wrong. But he failed to make
a strong case to his opponents that they are wrong. At the end of the
show his efforts only achieved a deeper divide between his fans and
foes, and were therefore utterly unproductive.
The silliness of Dawkins's debating strategy is that he tells his opponents they are irrational, and then proceeds to construct a rational argument that they are so. By virtue of its assumptions, such an argument is bound to fall on death ears. Were the opponents indeed irrational there is no way they can accept a very sophisticated rational argument. Were the opponents not irrational, they will just feel insulted from the outset, therefore too angry to agree with him even though if they thought he were right.
The larger point that Dawkins and others of his ilk fails to understand is that most people are not creatures of logic but of emotion. The very existence and persistence of religion in times of scientific and technologic progress is a testament to that. Arguments against religion cannot be dry and rationalistic, but must appeal to the same emotions that religion appeals to, because these emotions drive most people. And there is no reason why pro-science arguments cannot appeal to the same emotions, and win the argument in that arena as well.
Another mistake is Dawkins's focus on the existence of God. There is such a wide range of meaning assigned to 'God', from the personal God of the Evangelicals to the most diffusive forms of pantheism, that the debate can indeed shift rather to whether the concept has any meaning at all. But that is such a abstruse philosophical debate that it is silly to believe most people would get really excited by it.
More Emer action!
We have a new baby girl! Emer Alexandra Ghica was born at 8 am on
October 27.
Bertrand Russell said in a book (I forget which one) that the verb to be is an embarrassment for the English language because it conflates several related but distinct notions: existence, identity, attribution of properties, location and maybe others. Anselm's ontological argument for the existence of God is based on this confusion, which is obviously subtle enough to have caused Russell to briefly accept the validity of the argument before understanding its fallacy.
Another similar confusion is caused by the verb to play, although not of the same logical magnitude, because it conflates competitive play (chess is an example) and non-competitive play (the kind that puppies enjoy, say). Unlike the confusion surrounding to be, which is I suppose found in all European languages, the imprecision of to play is truly surprising because in other languages, Romanian for example, is avoided. In Romanian there are two nouns, joc and joaca that correspond respectively to the two kinds of play. There are also two recognisable forms of the verb, juca and se juca corresponding to the two kinds of play. This is even more surprising considering that Romanian has a much poorer base vocabulary than to English.
This confusion is unfortunate because the two kinds of play are two opposite activities. One is defined by strict rules and pursues well-defined objectives. It makes sense to talk of strategy in such context. It can be used as a metaphor for the economy or war, as game theory does. You can have professional players of such games. The other, however, is essentially rule-free and objective-free. Its values are aesthetic, if it has any values at all, because it is ultimately self-indulgent.
I recommend the book Homo Ludens by Huizinga as a good example of what intriguing theories one can construct on the basis of such misunderstandings.
Meletus died recently. His goodbye was beautifully dignified:
I came into this world in an inarticulate scream; I would like to go out of it doing something better.
Full text here.
The Pope recently made an intricate
theological argument
on the relation between Faith and Reason, in which he seems to
(unsurprisingly) assert that Catholicism alone strikes a good balance
between the two.
At some point he quoted Byzantine Emperor Manuel II Paleologus who during the
siege of Constantinopole (cca. 1400) is believed to have accused Islam
of being an inherently violent faith:
Show me just what Mohammed brought that was new, and there you will find things only evil and inhuman, such as his command to spread by the sword the faith he preached.
Angry Muslims around the word protested the accusation by attacking churches and even shooting a nun in Somalia. Some clerics declared that the Pope and everyone who believes Islam is violent should be killed. It's the latest development in the Global War Against Irony.
The Pope also takes a shot at Reformation and Kant:
When Kant stated that he needed to set thinking aside in order to make room for faith, he carried [the Reformation] programme forward with a radicalism that the Reformers could never have foreseen. He thus anchored faith exclusively in practical reason, denying it access to reality as a whole.
The Reformers and the Kantians have yet to react to this insult.
According to some superficial research I
did on Google, one of the simplest eukaryotic cells, the slime mould
dictyostelids, has a
34 Megabase genome. One of the
most complex eukaryotic cells, the spirochete leptospire, has
a 4
Megabase genome. There seems to be a vast complexity gap between
the simplest eukaryotes and the most complex prokaryotes. I think this
is a challenge to the evolutionary explanation of the emergence
of complex life on Earth. Evolution explains quite neatly the
apparition of life itself, and of prokaryotic cells and their
development. Once eukaryotic cells appear it also explains quite
nicely the further development to more complex life. The gap between
unicellular and multiicellular organisms is not that hard to bridge,
as the most complex protists and the simplest organisms have similar
complexity in the genome (at
around 40 Megabase). For
reference, humans have a genome at
about 3,000
Megabase, but the increase of complexity along evolutionary lines
seems smooth enough. It's the difference in complexity between
prokaryotic and eukaryotic forms of life that is staggering. The
official explanation, which seems to be endosymbiosis, is plausible
enough, except for two facts that I cannot easily square with the theory:
although there are some explanations, the mind-bendingly complex processes of mitosis or meiosis evolved without (?) any other simpler ancestors surviving.
most organelles are found in most eukaryotic cells. I would have expected endosymbiosis to lead to a greater diversity of structure within the cell.
Of course, I have little patience for those who cheaply exploit such interesting questions to further the cause of creationism or its equally intellectually bankrupt cousin "Intelligent Design". But here is another interesting article on the origin of eukaryotes. Fascinating stuff indeed.
My paper Geometry of Synthesis: A structural approach to VLSI design was accepted at POPL 2007. It describes a new technique for VLSI design that allows the synthesis of digital circuit specifications from generic higher-order functional programming languages with imperative features. It uses a semantic model inspired by game semantics and Geometry of Interaction, which can be expressed directly as a certain class of digital circuits. I will post the paper on the web only after the conference.
A phrase that is often-used in model checking is the state-explosion problem. It refers to the fact that the finite-state representation of a computer program requires an exponential number of states. Calling it "a problem" is quite amusing, because it entails the (false) hope of a solution. It is a bit like trying to build an engine and referring to the First Law of Thermodynamics as "the energy conservation problem". These are both inherent, unavoidable features of the phenomena at hand.
Much research effort is spent "dealing with" this problem, which is a rather ambiguous notion, which at least contains a hidden admission that "the problem" is not solvable. The unspoken word describing such efforts is "heuristics". Roughly speaking, a heuristic is a can-do attitude regarding a hard problem: let's try something anyway and be thankful that, in many cases, it actually works. I am not sure why the word itself, "heuristics" is not much used in the automated verification world. For example, the book "Model Checking" by Clarke et.al. doesn't use the word even once, preferring the more ambiguous "technique".
Heuristic approaches to model checking, especially abstraction refinement have been indeed quite successful, but why exactly it's not clear. It must have something to do with the fact that complexity bounds are stated on the worst case (see my earlier post of 09.09.2005) whereas in practice the worst case may or may not happen. For more on the incredibly difficult and deep problem of average-case complexity I recommend Impagliazzo's highly-readable introduction.
The War on Terror is being renamed again, and The Fight Against Islamic Fascism is emerging as the top candidate. But why pick on "Islamic" fascim? Are the other kinds of fascism excusable? Not worth fighting against? The Secretary of Marketing and Propaganda needs to think harder. The world cannot handle another failed rebranding exercise, like the unimaginative Global Struggle Against Extremism of 2005.
Meet my 2 months old niece, Isabella "Izzy" Campobasso. She is much cuter than your average baby. She is certainly much more fidgety.
The World Cup flattered to deceive. After a rocking first two rounds
of games it settled into the usual pattern of dull, defensive play.
Until the final, that is. Zidane's last hurrah was fitting for a great player. Gheorghe Hagi ended his international career on a red card for breaking Conti's leg in at Euro 2000. And his international club career on a red card in the UEFA cup final, for no some minor infraction. And his domestic club career on a red card for spitting at the ref.
Zidane didn't make it a hat trick of red cards, but his head butt on Materazzi was memorable by location, timing and, especially, delivery. None of that girlie shoving and slapping, followed by comical fake agony of the victim, stuff that gives football a bad name. No, this was unbridled violence, delivered with panache, skill and commitment. This butt would not be out of place in pro wrestling or ninja movies. It gave the victim sound reason to express agony, although Materazzi seemed too genuinely stunned by the attack to add to the effect.
Even in losing the final (and the plot) Zidane managed to steal the headline with his skills. Amazing!
The Weltmeisterschaft is upon us, and this is the best tournament I can remember, with plenty of cracking matches and cracking goals.
But this is missing the point. The World Cup has longed ceased
to be a sporting event and is now mainly a corporate advertising
extravaganza. One story with a higher-than-usual quotient of absurdity
is this
one: a group of Dutch fans were forced by stadium stewards to
remove their trousers because the trousers were advertising the beer
Bavaria, which is not an official WC sponsor. If this was supposed to
protect the official beer sponsor (which I shall not name) then it
misfired badly by giving Bavaria free publicity in media
around the world and close-to cult status and making of the said
trousers hot collector items. Brilliant!
Just got back from attending MFPS XXII in Genoa, Italy. Josh Berdine
and Richard Bornat appear in the picture, as well as a church door
with a sign (not seen) in latin promising eternal indulgence every
day. Imagine the possibilities.
Just got back from visiting a parallel universe.
Yes, the Boro game... how painful was that! Steaua won the first leg 1-0 and went 2-0 up in the return leg only to lose 4-2 an fail to make it to the final. It was the most outrageous collapse I have ever seen in a football match.
FC Steaua beats Rapid Bucharest to a UEFA cup semifinal spot. They will play Middlesboro next.
Just returned from a fine ski trip in Ischgl, Austria. What a hidden gem of a resort! I was skiing in knee-deep virgin powder in late March - enough said. [mov, pics]
America is exporting Liberty, perhaps to fill the market niche vacated by the Soviet Union stopping exporting the Revolution. Here's W making an ironic/disingenuous comment about the ungrateful reaction of the world:
I find it fascinating to listen to the voices from around the world as to whether or not it is a noble purpose to spread liberty around the world. G.W. Bush
Actually this is a bit unfair to the Bolshevics. Stalin's approach seems more sedate in contrast:
Every country, should it so desire, will itself achieve its own revolution, and if ti does not desire it, there will be no revolution. [From the Howard interview, 1936]
I hope there will be a happier ending to the American enterprise, but I am pessimistic.
A funny surrealist little game
I just returned from GEOCAL, the Geometry of Computation meeting in
Luminy, Marseille. I didn't know much about Geometry of Interaction but a personal
tutorial by Olivier Laurent focusing on the operational aspects cleared some
of the mystery. GoI seems very similar to a kind of game semantics which
makes no assumptions on the behaviour of the environment, so it is
not actually a model of linear logic. It provides a correct interpretation
for closed terms, because there is no environment: every module has been
plugged into properly defined contexts. The most amusing/outrageous moment
was JY Girard's classification of logicians as either "bureaucrats" or
"small-time crooks." On the touristy side, I took a longish hike, with
Cristiano, down to the gorgeous Calanques near Luminy. I also must mention
a dinner at Le Cuisineur, the best purveyor of pieds-paquets in Marseille.
My search for a Research Fellow came to an end with the appointment of Adam Bakewell. He has not much experience with game semantics, but his background in analysis and verification more than makes up for it. He also knows quite a lot about graph rewriting, something that I hope will come in handy in nice and unexpected ways. I sense our collaboration will be a very fruitful one. I was lucky to have three very good candidates, it seemed unfair to have to choose. Now I know Sven-Goran's agony when he must select a left-winger.
A good quotation:
President Washington, President Lincoln, President Wilson, President Roosevelt have all authorized electronic surveillance on a far broader scale. Alberto R. Gonzales. 80th and current Attorney General of the United States
The new year started quite busy, but maybe not as busy as last year ended. I was happy to find out that my paper with Andrzej was accepted at TACAS'05 (Compositional Model Extraction for Higher-Order Concurrent Programs) and a paper with Aleks and Ranko (A Counterexample-Guided Refinement Tool for Open Procedural Programs) was accepted at SPIN'05. I also gave my abstraction-refinement talk at Cambridge, after which I will give it in Nottingham and at GEOCAL in Marseille. This must be the work I have presented the most, but I think it's worth it.
A good quotation:
They have computers, and they may have other weapons of mass destruction. Janet Reno, 78th Attorney General of the United States.
2005 ended with two interesting trips, one to Israel for a workshop on
applications of semantic methods to software verification and one to Romania
for the holidays. Tel Aviv was a very pleasant surprise. Because of the
political situation I expected the city to be somber, but it felt young, bohemian
and bursting with energy. A memorable day-trip to Jerusalem rounded up the
touristic side of the workshop.
On the scientific side, the workshop was also a success. It was very well
attended and I was happy to meet the two great Ornas of model checking
(Grunberg and Kupferman). I was also happy to hear Pasquale Malacaria's talk on
entropy in programs, a very creative piece of work.
The picture is taken on the Olive Mount: Ooge de Moor, me, Samson Abramsky, Pasquale Malacaria, Alexander Rabinovich and Mooly Sagiv (squatting). The Dome of the Rock is visible between Pasquale and Alex.
back from seoul, korea. lovely trip, lovely food, lovely people. i discussed a lot of abstract interpretation with hongseok yang, my host, who is becoming quite an expert in the area. it was extremely useful because it helps sorting the wheat from the chaff, which is very hard to do by slogging through the literature.
have you heard of the scientific theory of intelligent design (i.d.)? it is making a stir
in the u.s., even bush spoke up in its favour. its point is that some features of life,
especially at the level of the cell, are so complex that it is more likely that they were
designed that they evolved. this point is argued quite skilfully, and i.d. can get some
mileage from the argument because at the level of cells there is a very weak fossil
record to witness evolution. of course, this point is rather philosophical and the
'science' built around it is fraudulent. it is, of course, a desperate attempt to sneak
creationism into school curricula.
what i find comical is that i.d. proponents, mostly militant christians, do not understand that i.d. as its stands does little to reconcile science with christian theology. i.d. professes to accept the scientific method, so it does not deny evolution as such because for multi-cellular organisms there is a irrefutable fossil record of evolution. only the cellular level is vulnerable to their argument. so according to i.d. there was a creator of the universe, but according to the fragment of evolution they cannot reject the creator of the universe was not the creator of man. man evolved. there exist a the fossil record to back up this theory. a theology consistent with i.d. is intriguing, because it implies a god creator of the universe, in which man was an unintended side-effect. this is is a fascinating demiurgic philosophy but at odds with christian theology.
asymptotic analysis of array algorithms actually makes no sense. the defining feature of array-ness is constant time access to the data. however, if we want to analyse an algorithm asymptotically -- in the size of the array -- then we need arbitrarily large integers to function as indices. and of course arbitrarily large integers can be read in logarithmic time at best. so even if we are ready to accept the fiction that random access in arbitrarily large memories can be done in constant time, the index itself cannot be processed so. the array is a data structure that does not even make sense asymptotically, because for large enough sizes it ceases to be an array.
more travel, with stops in frauenchiemsee for the appsem ii workshop
and
lisbon where i visited claudio hermida. the workshop was situated in
the picuture-perfect, if a little dull, setting of a convent on an island
on a lake at the foot of the bavarian alps. i greatly enjoyed a talk by
and talking to joe stoy from bluespec. they are doing some really
interesting stuff with hardware synthesis from functional languages. the
lisbon trip was very good,
very productive. i always enjoy talking to category theorists because it
sometimes turns out some ideas that they think are pretty trivial can be
helpful in practice. especially with games, which are so mathematically
unwieldy, it is always a huge bonus, a lot simpler, if you can prove
something, anything, from first principles rather than in the concrete. i
think the work in our sas 2005 paper stands to benefit substantially from
some of claudio's insights.
The picture shows me diving into the Chiemsee. Peter O'Hearn is sitting on the dock.
i just returned from the static analysis symposium in london, where i presented joint work with ranko and aleks on data abstraction refinement for games. the conference was enjoyable and well organised. the work presented by byron cook, on termination analysis stood out. it is the third talk in this line of work i am attending (andreas podelski at lics'04 and andry rybalchenko at popl'05) and i am increasingly convinced it's important work which i should understand much better.
as i get more interested in 'practical' software verification i am struck by how inadequate is our commonly accepted measure of goodness for algorithms, asymptotic complexity. in practice the asymptotic complexity of an algorithm tells us surprisingly little about the usefulness of an algorithm. it is indeed rarely the case that polynomial algorithms are useless, because huge constant factors or powers involved, but it happens. however, exponential algorithms, or other algorithms deemed computationally intractable by their complexity class, can often be useful. even semi-algorithms can often be useful, but asymptotic complexity cannot really yield a meaningful characterisation. the trick is though that not all putatively intractable algorithms and not all semi-algorithms are useless. sometimes the worse-case behaviour kicks in only rarely and the algorithm is otherwise fast. sometimes the algorithms is uniformly inefficient. semi-algorithms may diverge often or rarely, and when they don't diverge they may be fast or slow. average-case analysis might work for algorithms, although it's very hard to determine it, but how do you compare semi-algorithms? i am not aware of much work in this area. the notion of improvement by david sands, and the notion of relative completeness by ball, podelski and rajamani seem very good steps in the right direction. unfortunately most researchers who develop high-complexity algorithms or semi-algorithms seem happy to "test" their programs against carefully chosen sets of examples, which i find utterly unconvincing.
in the wake of hurricane katrina, the mighty u.s. of a. looks just about as solid as its former superpower rival, the soviet union, circa 1988. the soviets got bankrupted by extravagantly expensive efforts to export the revolution to peoples who weren't really interested, while their own country was falling apart. it's astonishing how the americans learned nothing, and are doing exactly the same. they also seem to employ the same kind of brazen flatulent black-is-white propaganda, with ever reduced effectiveness on an increasingly immunized and skeptical citizenry.
i don't understand ikea's success. i bought some stuff from them last weekend, only the second time i ever bought ikea merchandise. the prices seem good when you see the items, which are often quite pretty, on display in the showroom but once you get home and you notice the poor functional design and the shoddy execution it looks like a rip off. the chest of drawers "vestby" which i bought for the attic was desperately difficult to assemble. the assembly instructions seem to defy the laws of physics. the drawers are formed by a sheet of plastic that must be folded along 10 lines. the problem with folding plastic is that, unlike in the assembly diagrams, it does not stay folded. as soon as you let go it reverts to the initial flat configuration. you must use your hands and feet very creatively to hold the thing in place before you can snap the handles on, which fixes it into place.
which you cannot do at all if you folded the drawer imperfectly because the handle holes must align perfectly. the chest was also difficult to put together. you need to first fit the nuts through some holes and align them blindly with the screws coming though different holes. then you need to tighten the screws using the hilarious little toy screwdriver provided with the package, which needs to be taken out and readjusted after every half-turn. it takes more than 20 turns to tighten one screw so i was lucky to have a real screwdriver. the chest also sets the classic trap of providing three almost identical panels which are not actually identical. assembling them in the wrong order is only noticeable at the very end and requires disassembly. the quality of the item was deplorable. some of the wood was cracked and some of the holes were obstructed by debris or did not align properly. never again.
it seems reasonable that since i have quite a few friends scattered all over the world it's more efficient to keep an on-line log of what i've been up to rather than send a large number of sligtly modified emails whenever i want to report something. this seems to be a good time to start since we just moved into our new great little house, after struggling to accommodate into our previous house which was too big, too decrepit, too expensive and too far away from where we work. next month will be quite eventful, i think, because there are a number of work-related trips coming up.