Friday, October 29, 2004
I had a long discussion with Annette about this spot and whether the coworkers liked the idea or just nodded because they didn't want to ruin their bosses succes of coming up with something they thought was stupid. The spot was on the internet as well but now you have to pay 15 pounds to see it, so unfortunately, you cannot judge for yourself.
I tell you about this because not being able to read people's expressions is one of the symptoms of autism. Not that I claim I am autistic, but sometimes you wonder whether being a string theorist brings you quite some way in that direction.
I enjoyed a lot the 'curious incident of the dog in the night-time written from an autistic boy's perspective. It is alarming how he tyranises everybody around him without noticinig. A must read! And the chapters are numbered from P, the set of prime numbers rather than N.
There is also Asperger's syndrome, something like Autism Light. That is probably pretty common in certain academic departments. Reading this page makes clear that I and most people are far from that stage (there was a student in Cambridge who was a diagnosed sufferer of asperger's syndrom and knowing here makes clear that all the nerds and geeks around you are way in the middle of the bell curve) but sometimes you wonder...
If you want to see for yourself, wired magazine has an online test. They say, the average person scores 16.4 and people with asperger or mild autism typically score 32 and more. My score was 28. What's yours?
Thursday, October 28, 2004
Update: Unfortunately, the above link now forwards to the general Renault website and the clip appears to be gone.
Second update: YouTube helps out:
Here is the one with Animal:
which I was talking about. There is another one here:
Wednesday, October 27, 2004
As a souvenir from that workshop I brought with me a book with proceedings. That is not so unusual, but in this case these were the proceedings of the Nobel symposium on Particle Physics in 1986. The conference name however is not telling the truth, it was more about strings than about particles. It is really amusing and amazing to read in this volume. Obviously, this was before the second string revolution of the 90s but there are some hint already around.
It starts out with a picture of all participants (18 years from today). Once I found a scanner I will let you participate in this picture, so far I can tell you that I am able to recognize the following people (in no particular order): Green, Van Nieuvenhuizen, Schwarz, Gross, Witten, 't Hooft, Ellis, Hawking, Salam, de Wit, and Fiedan. The list of participants is much longer but the picture is not of supreme quality.
The first contribution is a review by Michael Green that starts out with an expose of Regge behaviour of scattering amplitudes (much more detailed than chapter one of GSW) and then covers all the 'recent' developments. Michael finishes by listing 245 references which costed him quite some efford as he told me. Then comes Olive about KM algebras and Alvarez-Gaume about anomalies and index theorems, Gross explains the heterotic string (mainly text, very few formulas). Many more contributions follow, Witten talks about string field theory as do many others.
In the end, it gets particulary interesting, when they predict what will happen next. John Schwarz' contribution has the title 'The Furture of String Theory'. So let's see, what he predicts: He lists six approaches to string theory: 1) operator methods a la Olive, 2) Polyakov's path integral, 3) 2d CFT by Friedan and friends, 4) 10d effective action 5) light cone gauge field theory 6) string field theory and he predicts "All six approaches wil play an important role in the future developments". OK, he got that mainly right.
Then he talks about background independence, but I guess that will have to waiut for the next twenty years. His next prediction is that type II and heterotic will be shown to be finite at all orders of perturbation theory. Thank's to Berkovits, he got this right as well. His next prediction is "World sheets of infinite genus will plau an important role in future nonperturbative string theories.": That was only half fulfilled: Nonperturbative methods were at the heart of the recent developments, that's right, but I don't know anybody who thinks about those in terms of infinite genus worldsheets.
The next point is the uniqueness of the theory. He quotes Harvey who said that "various heterotic string theories are really different versions of the same theory" and then predicts "It would be most satisfying if there were only one string theory, which is the correct microscopic theory of nature...Presumably the heterotic theory shoud be the survivor. If this is going to happen the type I and type II superstrings would have to be found to be either inconsistent or equivalent to the heterotic theory". Well, M(other) decided for option 2.
The next prediction is an easy shot: "Future progress in string theory will be accompanied by significant advances in mathematics". And the last one is probably the best one: "String theory will be an even more vital and active subject at the turn of the century than it is today ".
The conference summary is done by Gell-Mann. Among many other things that I do not have time to report on right now, he mentions Luis Alvarez-Gaume as giving the lecture that should get the sublimal award for the meeting for flashing transparencies at the highest rate. He reproduces a page from his notes (I should scan that as well), that reads
picture of worldsheet ----> some flat diagram, phi=integral dphi not single valued...IMPORTANT....???...1) phi --> phi+const. 2)------ [two more transparencies]...??...Assume winding of soliton around a-cycles and b-cycles ==> 3rd term... 2nd term computated in holonomy... 1st term??? ...S=S1+S2+S3
And he mentions S-duality of type I and Het-SO(32): "Ther is even the half-joking suggestion of Ed Witten that the two SO(32) theories might be physically equivalent, is such a way that the dilaton field phi is one formulation is related to phi^(-1) [should be -phi] in the other." You can find this conference on spires
The following years, it became the rule that you had to queue for a train ticket for at least ten minutes, much more in peak hours. Then they introduced vending machines where you could buy your ticket with your credit or debit card. But those were restricted to 'simple' tickets. For more complicated things like reservation of a seat with a socket for the laptop or trips with complicated stops in the middle you still had to queue.
But now, you can surf to the above address, register your name etc. and then search for a connection and get your ticket as a pdf file. This pdf document contains your name, part of your credit card number and a 13 character key containing letters and digits. On the train, the conductor slides the credit card through some gigantic PDA, checks a photo ID and types in the security key to verify the ticket. So far, that sounds reasonable.
On last weekend's trip to Berlin, I had to learn hat a fourth step is important as well: The conductor insists on stamping the ticket. Before leaving, stupid me had been so clever to move the printout from my bag to a pocket. However, this is not an atomic procedure and multitasking interrupted right in the middle. Effect: I left the printout on my desk. I still had my laptop with me that had the pdf on the harddisk and I had it as well on my USB stick. But at Bremen central station I couldn't find a place that could print out one page of pdf for me! I had to buy a new ticket because the cannot stamp my laptop's display. I can return the online ticket, but this costs 15 Euros. Why?
For old style tickets it makes sense to devalidate them by stamping them. But why do they insist on this once they gave me the pdf that allows me to produce hundrets of printouts of the ticket? It seems, this rule procedure was implemented by somebody lacking some basic understanding of how computers work.
Tuesday, October 26, 2004
Under Win2k, the problem is probably that I have not yet found out which encryption key to use. That problem most likely has an easy solution once I seriously investigate it.
As we all know, Win sucks and GNU Linux is much cooler. Thus I resized the partion (took me a day due to some sectors marked damaged by some hard disk driver) and installed debian. With the help of Christian I even manged to compile and install a 2.6.8 kernel. And with some heavy googling I resolved an interrupt conflict.
Some things that still need some fiddling: Make the irda port recognize my new Sony Ericson T630 mobile phone and transfer image and sound files. Get wireless networking to work (the experimental driver module for the intel 2100 wlan chip is supposed to come in a .deb package that cannot be loaded from the usual debian servers...), migrate mail reading/composing to the laptop (I still use pine running on a machine in Cambridge). Of course just using IMAP to transfer the mail to the laptop is not the hard bit, getting spamassasin do an equally good job as the spam filter in Cambridge is much harder and finally I want become as versatile in mutt than I am currently in pine. I still find that UI very confusing and it does not work properly for me (for example sorting mail semiautomatically in my 100+ mail folders). You see, I am quite behind. And there is also some physics to do....
I find this hard to believe. From what I can see a typical poll has a (statistic and systematic) uncertaincy of at least 2-3% (absolute, not relative). This means that in many states it is in principle impossible to make reliable predictions and due to the very non-linear "the winner takes it all" system, these uncertainties become amplified when it comes to the majority in the electoral college.
From 2000, we still remember that it is not the total number of votes (assuming there are no further problems in deciding which ones are valid) that determines the next president but also how they are distributed over the country. And none of the reports on recent polls that I have seen comment on this. They all state only that Bush is leading by a couple of percent but not how that translates into votes in the electoral college.
Talking of conspiracy theories, there is still the issue of Bush being radio controlled during the TV debates. I must say, I would prefer if he would have been. In the end, it is also his advisors that determine policy so why shouldn't they have their say in the debate? Bush by himself is probably not very powerful or influential, it is the people in the background that you never get to see because they are hiding in their think tanks and operate the teleprompter or the RC for the president's mouth from there. And they make sure that groups that put huge sums of money into getting candidate x elected get their money's worth of their investment.
The question is: Do I really need the "linear combinations of" or can I get away without it. Your task is to find an example of a Lie algebra, where some
[a1,b1]+[a2,b2] cannot be written as [a,b]. Can you characterise algebras where this happens?
BTW, this is an exercise for you, I think we managed to figure out the solution. I just post it as I like it because it is not a standard exercise of the form "Show that the following assertion is true using the standard methods explained in class", it rather requires more creativity to solve it.
PS: Jacques Distler is right, TeX in a blog would be nice!