> The paperback edition of Brad Stone's The Everything Store: Jeff Bezos and the Age of Amazon — a book Amazon disliked so much it denounced it — is suddenly listed as 'unavailable.

This would be disturbing if true.

I just did a search, All editions are available and in fact I just bought the Kindle edition now.

Sorry I can't upvote you as I have insufficient karma..

> This is just more of the same delusion. Even a simple, inexpensive DAC

The oversampling that DACs use is to avoid artifacts from the reconstruction process. I have implemented these algorithms myself. But the DAC cannot reconstruct what is not there. And there are inevitable losses in the digitization process with a limited number of bits.

> 16 bit isn't imperfect given that it scales from the lowest detectable audio level to a level that would cause hearing loss.

You are confusing two things here. The 16 bits has to suffice both for the dynamic range of the music and for the structure of the wave form.

The dynamic range of the human ear is ~100db. 10db is worth about 3.32 bits. So 100db is about 33 bits, or twice the entire space we have already.

On top of this you need room for the wave form. Realistically if you want a pathetic 8 bits for the wave form, you are left with 8 bits for dynamic range or less than 30db.

So much bro science in this thread.

Citation required.

This a a chronic problem with music such as classical music which has extremes of dynamic range.

> And the irony, the irony. A NYT journalist - from the home of the liberal arts graduate - lecturing tech people how they should spend their lives.

Even though the actual author is a CS guy, I stand by the comment about the NYT.

> No it assumes there is no interesting signal above half the sampling rate

The **theorem** assumes **no signal at all** above the cutoff.

To ensure that there is no signal above the cutoff you have to filter out signal above the cutoff. And any filter has a slope, it is not a cliff face. You are going to lose some signal in the audible range from the filter.

This is in the ADC phase. All the talk about oversampling etc in the DAC phase is beside the point. The signal has been lost in the ADC phase, the DAC can do nothing about that. The oversampling in the DAC is to remove artifacts from the DAC process. I have actually implemented these algorithms myself.

Absolutely. I see this as part of what you have to offer, or otherwise make the money so compelling that it overrides these factors.

When people complain "supply is less than demand", they often forget to include "at a price I want to pay". It's funny that you often hear this complaint from people who pose as champions of free enterprise.

Yes there is a shortage of technically competent people prepared to work 60 hours a week for minimum wage.

If you pay the market clearing price, you will not find a shortage.

> do you hold

do not hold

> If I had a nickel for every time an audiophile tried to explain...

I used to share this view but eventually I concluded that "CD Quality" is not as good as it gets.

This is the classic ludic fallacy that nerds are prone to - confusing theory with reality. http://en.wikipedia.org/wiki/L...

Nyquist's theorem has some assumptions that do you hold in the real world and which actually impact real world sound quality.

1. It assumes there is no signal above the cutoff (1/2 the sampling rate). If this assumption is not met ie in the real world, then annoying 'aliases' appear in the sampled signal. To fix this, you have to have a low-pass filter. The low-pass filter, by its nature (physics) has to start cutting out signal well below the theoretical cut0off. So there is inevitable loss of signal well under the cutoff.

2. It assumes perfect, 100% accurate samples and reconstruction. Instead we have imperfect 16 bit resolution samples and heuristic sampling and playback. Reconstructing a good playback signal is a bit of an art. The main impact is the loss of dynamic range. Engineers are forced to limit the dynamic range of the music to avoid excessive loss of accuracy and/or clipping.

I am not a golden-ears person myself but I have friends who are, and gradually they have convinced me that there is a real loss from 16 bit 44kHz samples versys vinyl. I find mp3s unlistenable. Flac and also implicit higher sample rates on DVDs I find OK. I like the lack of noise on digital recordings (no tape hiss or surface noise). But I would happily replace my CDs / flac with higher resolution sound.

The problem is a generation brought up on mp3s expects more of the same.

This.

Society treats actual researchers like s**t. Years scraping by on one tenuous post-doc after another, and that's after 12+ years qualifying for the job, accumulating debt and then living on tiny graduate scholarships.

You get what you pay for, America.

And the irony, the irony. A NYT journalist - from the home of the liberal arts graduate - lecturing tech people how they should spend their lives.

> long-term

I see the subtle flaw in your thinking. Anything that happens post the current CEO's tenure is discounted to zero. If doing this would cause the earth to fall into the Sun in 10 years, it would still happen.

> such a crash-ridden heap of bugs.

Luxury. My rich cousins have that release. Ours was much worse, perhaps because we live in a hole in the road.

On my Android it takes 20 seconds to display a page on the pdf file I am reading at the moment. 2014!

... and everyone becomes obsessed with politics and 'visibility' - because those who do not are out the door.

People have been trying to create better programming languages for 60 years. If there were any great and beautiful solutions they would probably have been found.

Domain specific languages have been useful but the problem is that there are many domains. Once you know the DSL it's good but there is the overhead of learning the DSL each time. And what if your requirement is just a little bit beyond what the DSL supports? Back to the good ol' 3GL.

Lisp (and to some extent Ruby) offer the possibility to create your own DSL. Once you get used to the parentheses, there is very little syntax to learn with Lisp. Typically in a project in Lisp you create a DSL for the problem and then write the solution in the DSL (see Paul Graham's book "On Lisp" available for free). This makes for fewer lines of code but more effort to understand per line of code, compared to 3GLs.