I often use Youtube to check out a track before purchase. For this and publicity if's fine, but if you can seriously enjoy listening to a track on youtube through a decent (note, not audiophile) hifi, you need help.

And if the stock crashes, he's landed with a massive loan and no way to pay it back. Thus bankruptcy.

I recall hearing about mark-to-market in a documentary about the Enron disaster. Think about this though: if I hold $500m of shares in a company at today's valuation, then next year they're worth $2b, and the following year they are down to $500m again, should I pay tax on the 'gain' in between? (approx $500 say) and then do I get that tax back when the shares go down? Without care, mark-to-market taxation would be a tax on volatility of shares, rather than one on real value. As for borrowing using shares as collateral, maybe there should be some tax there, but I can still see problems.

The classic in English is the ough ending. Count how many different ways it can be prounouced. (Oooh, Uff, Ow, Oh, Uh... any more? And I have no explantion for a non-native speaker as to how that came about.)

And on the UK Mac keyboard, you then have to use option-3 to get a hash(#), whereas typical UK keyboards have dedicated keys for both. Makes programming certain languages on the Mac a little tedious (having to use a shift-key combination everytime you want to comment in e.g. PERL or Ruby).

Computer Music is a refreshing exception. There's enough free plugins and samples for a beginner to get started, and a video by a major producer in their studio illustrating one of their tracks production. CM are generally the only cover DVDs I go for and hang onto.

Computer Music, Digital Camera (Future, via MyFavouriteMagazines.co.uk), Sound on Sound (though not currently) via their website, Linux Magazine (not current) via ads in their magazine, Amateur Photographer (via their website). Currently trying to kick my sub habit, or at least get it under control. That said, I am UK based, so I'm not sure what its like subscribing to UK magazines from overseas.

Better delete that crypto volume ;-)

Time frequency uncertainty is a known issue in signal processing, and an explanation can be found early on in A Wavelet Tour of Signal Processing, for example. The post you are replying to stated that this is a fundamental property of the discrete Fourier transform, which it is.

To be able to personally tailor medicine to an individual, you need an effective diagnostic mechanism, and an effective way to decide, based on the diagnosis and the patient, what course of treatment will work. Trying to do this on the basis of a patient's genetics is unlikely to be effective in the near future, if at all.

Te begin with, studies of the kind we see today tend to give results like 'gene X affects incidence of disease Y by n%'. To rely on this for a diagnosis and treatment amounts to a guessing game, and the number of such n% guesses compounded together will cause accuracy of the diagnosis to be little better than random chance, yet appear to have the certain blessing of the medical establishment. Establishing the effect of a gene is, in any case, far less certain than seems to be made out, because there is little understanding of how an altered gene causes a problem even if a correlation is detected.

I fear personalised medicine is the road to mass Russian Roulette medicine, and I hope the 'brave new and shiny' factor doesn't cause it to be overly relied upon.