<t>You can only be closer to an understanding of the C++ standards process than I am. No reason to expect that illustrious body to move quickly though ;-)It's worth having a skim through the paper. I was a poor number theorist but delinearisation might just be my new favourite word! The TRNG library...

<r>The issues of internal structure of the sequences produced by pseudo-random number generators which become more pressing when using multiple streams from the same generator for the same calcualtion are discussed well here:<URL url="http://arxiv.org/PS_cache/cond-mat/pdf/0609/0609584v2.pdfWith"><L...

<r>You need P types to break the rules (challenge if they're also E, otherwise just get on with the breaking if they're I). There seem to be a lot of J's here, is this generally true of banking, and if so, does it explain the current mess??E/I NN T/F PPPP <E>:-)</E>(It is a fact of life forever stra...

<t>I've used mapping techniques extensively facilitating group design exercises for such exciting things as tax office internal workflow projects, as well as risk systems in banks. The key difference between mapping a conversation and the more popular "mindmap" is the graph which results. Real conve...

Well work hasn't gotten any less demanding. Is there random beer to be had?

<t>Moving from coding parallel numerical calculations to multi-threaded applications can be a rude shock. Asynchronicity is much more complex than the scatter-gather style of OMP loop-level parallelisation. To what extent do finance codes need true multi-threadedness? Where I work we have enough pri...

I used to work in technology commercialisation. Software can be a tough game at the best of times. Difficult to charge enough upfront, most of the money is in maintenance agreements and the like. Also, I would be very surprised if anyone would trust a black box at the moment

<r>QuoteOriginally posted by: katastrofaQuoteOriginally posted by: riskhommeMaybe this is a question for Paul: where are the fully numerical methods that engineers and scientists use everyday??These methods rely heavily on the fact that they model systems of 10^23 identical particles. Back to the "l...

<t>QuoteMathematical finance should look a lot more like molecular dynamics than statistical thermodynamics.Absolutely! Two points about water:1. The first models could not produce a finite pH, the next generation were considered a huge success with pH = 35, third generation got down to 13...sound f...

The Fed has asked banks to submit a list of top 20 exposures. So in principle they will at least start to know who is exposed to whom. In the bank I work at the P&L of swaps has fluctuated by over a billion in the last month. Wild ride indeed.

Is this happening? Any clues as to where?