Tuesday, June 16, 2009

Silence

I have not had time for blog postings. My response to the Mahndisa in earlier posting gives the reason why and also some ideas about the recent situation in coupling constant evolution. I hope that I can write as summary within few days.

Dear Mahndisa,

numerical work makes me rather non-communicative;-). I have been working with the numerical realization for the model of coupling constant evolution based on quantum criticality. Numerical work is is not easy at this age and would require a hard wired brain at any age. To make challenge even more difficult, I am forced to use MATLAB without compiler and there are a lot of loops. The basic challenge is how to make things fast by making as much as possible analytically. This kind of problems sound of course childish and crackpottish in the ears of anyone allowed to use the computational resources of any physics department of any University. Perhaps I should be ashamed;-)!

There are also other reasons for being not so communicative. Last week MATLAB went completely mad: probably I became a victim of a clever virus attack, nothing standard against which I am well-shielded. This was not the first time.

Add to this jeremiad the problems of everyday life (for a week I have had nothing at my bank account and getting support social office is slow nowadays since there are long ques) and you begin to understand why the working conditions are not very inspiring. We are however in Finland and in the academic circles of this country thinkers are regarded next to criminals and the best manner to treat them has been found to be the academic equivalent of Siberia.

In any case I have made a lot of progress in understanding coupling constant evolution. The question whether the proposed realization of quantum criticality works is still open. In any case, at ultrahigh energies the behavior of em couplings strength would be like that for asymptotic free theory if criticality is accepted. For low energies the criticality is consistent with standard model behavior for fine structure constant (its value at electron and intermediate boson scale are the constraints). I do not yet know whether the low energy and high energy behaviors are consistent with each other or not. The calculations are desperately slow.

This problem led to the ask whether p-adicization of the theory is necessary to realize criticality. Within two days this led to a rather precise recipe for how to p-adicize the theory in terms of p-adic fractals- creatures which I discovered within first year of p-adic TGD but for which I have not found direct application in TGD hitherto.

The recipe was very simple: consider real Lorentz invariant amplitudes, map Lorentz invariant kinematic quantities to their p-adic counterparts by some variant of canonical identification to get p-adic calued functions with same functional form, carry out arithmetic operations such as the summation of perturbative contributions using p-adic arithmetics, and map the result back to reals to get a p-adic fractal.

You just go to p-adicity, perform arithmetics there and return to reality to see what you got! In this manner the difficulties related to p-adicization such as the non-existence of p-adic definite integral, and the problems with minus sign and imaginary unit can be circumvented and the outcome cannot differ too much from real physics prediction.

I hope that I can write about this within few days. The recent situation concerning bosonic emergence in quantum TGD framework given in the new chapter Quantum Field Theory Limit of TGD from Bosonic Emergence of "Towards M-matrix".

Best Regards,

Matti

8 Comments:

At 12:02 AM, Blogger Javier said...

Hi Matti, May be I could help you here.

A friend of mine, physicist working in informatic/translations has a 5 servers farmer and offered me the possibilitiy to run complex scientific software in it if I needed.

I am not sure whether matlab software can take advantage of parallel computing but if so maybe he culd run the matlab program in that system for you. Tell me if you are interested and I'll try to speak to him.

 
At 1:29 AM, Blogger Matti Pitkanen said...

Dear Javier,

thank you for a nice offer.

The programs involve input from screen during computation since basically experimentation with various options is in question. If basic modules could be compiled it could be enough to do everything using home computer. MATLAB compiler however costs and I am living with unemployment money and what social office can give to me.

I should be able to write and develop program modules at home computer, send them to the computer system, run them there and get the results via web connection. What do you think? Is this possible?

 
At 9:48 AM, Blogger Javier said...

Ok, I'll if something can be made.

I guess that that finding somone who could compile your source code wouldn't be too hard. I haven't used too much matlab (I prefer mathemathica and only use matlab because it is required for university work) and I didn't know that It could compie code.

In fact I am aware that there is a free version of a program totally cmpatible with matlab for the unix platform (that can be run in a virtual machine under windows, you don't even need to install linux). that, maybe, has the compilation utilty also..

Anyway, I'll seek possibilties and I'll tell you as soon as I know. But seemengly I guess that someway or another there will be a solution ofor your problem.

 
At 7:42 PM, Blogger Matti Pitkanen said...

Well, it would be better to find a millionaire who could afford MATLAB compler (costs less than 1000 euros;-).

This virtual version sounds interesting. It might well contain a compiler.

Thank your help.

 
At 12:29 AM, Blogger Javier said...

In fact millionaire financing freelance scientific are welcome for whatever purpose ;-).

To avoid possible mistakes I'll be specific. By "matlab for linux" I was talking about octave, that is a package that almost fully supports the matlab language (see http://www.gnu.org/software/octave/FAQ.html for a FAQ site).

In fact there is other matlab clon for linux, scilab (http://www.scilab.org/).

In fact, from the last time I had watched, have appeared windows versions of those programs. That means that you don't need a virtual mmachine to emulate a linux under windows to run linux versions anymore.

Anyway, what really matters is not to have a matlab clon, but the option to compile matlab code. Well, what I have found so far is not clear about that. Seemengly there are people triying to make such a compiler, but I am not sure of how effective it is.

This bring us again to matlab. I have watched and the version of matlab I have instaled on my computer hasn't the compiler so just now you just can't send me the source code. But don't worry too much about it. One one hand the next day I would go to my university I'll seek for a computer that actually has it installed. Surely there is one who has and I could compile the source code there.

But, of course, there are more alternatives. I find totally stupid that one couldn't do appropriate research because of this minor problem. Just give me some time ;-).

 
At 2:59 AM, Blogger Matti Pitkanen said...

Suppose it is possible to run compiled programs using MATLAB without compiler. If so and if you were able to find a computer with compiler from this planet then I could send the basic modules and you could compile them and send back.

Matti

 
At 3:32 PM, Blogger Mahndisa S. Rigmaiden said...

Matti:
Thanks for the response. You know I often think of Heisenberg and Schroedinger picstures of qm as bastard relatives of one another. They are related to one another (in non relativistic qm) by a very simple first order differential equation.

You said this: " The recipe was very simple: consider real Lorentz invariant amplitudes, map Lorentz invariant kinematic quantities to their p-adic counterparts by some variant of canonical identification to get p-adic calued functions with same functional form, carry out arithmetic operations such as the summation of perturbative contributions using p-adic arithmetics, and map the result back to reals to get a p-adic fractal.

You just go to p-adicity, perform arithmetics there and return to reality to see what you got! In this manner the difficulties related to p-adicization such as the non-existence of p-adic definite integral, and the problems with minus sign and imaginary unit can be circumvented and the outcome cannot differ too much from real physics prediction. "



So you were able to get over the difficulties of connecting these two pictures. Great! Where in your TGD books do you discuss this? I wish to read relevant chapters.

The p-adic mapping to more conventional picture seems rather daunting. I am glad you figured out a way to get around these difficulties.

 
At 9:59 PM, Blogger Matti Pitkanen said...

Dear Manhndisa,

I have not actually seen the relationship between Heisenberg and Schrödinger pictures as the problem. Problems are essentially number theoretical. And "ontological" as philosopher would say. I have a bundle of general ideas developed in various chapters of various books. I try to give impression about just the bare essentials.

Well I tried! The impression had too many characters (blog program does not accept impressions longer than 4,096 characters) and I decided to add it as a separate blog posting. It appears within hour or two.

The relevant text can be also found from the last section of the new chapter Quantum Field Theory Limit of TGD from Bosonic Emergence of "Towards M-matrix".

Matti

 

Post a Comment

<< Home