randombit

I'm also fond of

  "What's the thing you like least about your job?"

Tahoe-LAFS provides a storage service that is encrypted by the client, but the original uploader of the data is still able to delegate read-only or read-write access to others on a per-file or per-directory basis.

The overall design, including how it is possible to do secure delegation, is fairly well described in a paper from the 'Storage Security and Survivability 2008' workshop - http://tahoe-lafs.org/~zooko/lafs.pdf

I don't see the connection between your first sentence and your second.

I don't believe reality television adds anything of value to the world, but that doesn't mean I think it should be 'stopped' (and what does that even mean?). Can you explain?

However in this case the missing free lunch may be for anyone with an expectation of making revenue off ads, rather than the people who won't even notice when site X happens to not update too often anymore due to decline in ad revenue, because sites A, B, and C came along and filled that attention gap nicely thankyouverymuch. (And if people do actually care enough to notice and care if site X stops producing, then X doesn't need ad revenue to make content creation profitable)

I'm not quite on Pitcairn Island, but I live and work in rural Vermont, and by far the biggest problem with it is precisely the lack of a reliable internet connection. Right now for me, the only viable option is satellite, which is horrible (225 MiB/day download cap, 800 ms ping times). I'm currently investigating a T1 line, which will be expensive, but (slightly) cheaper than renting an office in an area with DSL access. (Moral of the story here is clearly to not fall in love with a girl who owns a house on the end of a dirt road...) I'm sure many (though certainly not all) other remote locations are as bad; so don't make assumptions on that before you pull the trigger on a specific area.

Another factor of living in the middle of nowhere, especially after living in DC and NYC, is the technical/social isolation is pretty intense. I'm about 10 miles from Dartmouth, and there is small tech culture there, but it's not like there are monthly Lisp or Scala or security meetups. Or even a student ACM chapter, for that matter. No matter how you slice it, it's hard to replace quality in-person interactions with online. IRC and mailing lists are great and useful things, but being able to hang out drinking coffee with someone while you sketch out ideas in a notebook is pretty damn important too. (There are plenty of interesting people out here, but most programmers, especially the good ones, will of course gravitate to urban areas, so the population of tech people out here is near zero).

Unfortunately they may have passed the Rubicon with that. I suspect that many of the people (including me) who would have been willing to pay have already gone ahead and signed up for another paid bookmarking site. The most active users are the ones who are most likely to have already heard about the shutdown, and the ones who would be most concerned about it.

Just as well, since it looks like pinboard gets the things that bugged me about delicious right.

> No one is having problems in C/++ because of this.

No one is having problems with this in C because C doesn't have exceptions.

No one is having problems with this in C++ because C++ supports RAII via destructors which are guaranteed to run on stack unwind.

Especially when you throw in going from Xt (!) to Cairo/GTK - you could probably save 50% of the line count right there.

Using an acyclic graph of chained hashes to represent change history has a long history in revision control - it was used in Monotone (which Linus considered as an option instead of writing git, but rejected as it was written in C++), in OpenCM, in Xanadu, and probably in at least half a dozen other systems I've never heard of.

Bad headline: the 'high-risk' defects include resource leaks and other problems that could cause crashes or data loss but are not necessarily exploitable.

And, of course, this is assuming either Coverity manually checked and confirmed each of these problems (possible, but not mentioned in the article) or that Coverity's scanner has a 0% false positive rate (highly unlikely to the point of patently absurd; my experience with static analyzers is false positives being 50 to 80% of the total though it's been a couple of years since I used Coverity specifically).

C++, CUDA, Python

Performance wise I've seen some good results from Clang SVN in the last few months. Not universally faster (or slower) than GCC or ICC, but I have seen (important, useful) cases where code was 30-50% faster compiled with Clang than anything else I tried. (I also saw cases that had terrible performance, and occasionally miscompiled code, though mostly it was +-10% of GCC).

I can't say I understand why anyone would care too much about compilation speed, though - GCC+ccache is about as fast as Clang+ccache, as the primary bottleneck is disk throughput.

In 2008 one of the DARPA 'Young Faculty Awards' went to someone at UW-Madison for a project named "Super Artificial Eyes"; I don't know nearly enough biomedical engineering to parse even half of his paper titles, but it appears he's still working in this area. I'm sure the military would be very interested in this and similar techniques, thus the DARPA funding. I would guess that for a long while after it's actually viable it would remain mostly an expensive toy for nation-states but eventually it would leak out to people who can make use of it (sports stars, for instance) or people who are just into self-modification, long before it becomes common among normal humans.

How will people react to it? Just ask yourself how older people reacted to electricity, the Internet, genetic engineering of infants, &c. It will seem normal once enough people have grown up with it, for the rest of us (outside of the neophiles, which probably is a majority on HN but a tiny minority in the general population) it will just seem crazy and strange.

Er http://www.ksplice.com/

"When the density of matter reaches gargantuan proportions (more than about 1,050 kilograms per cubic meter)"

I assume there is an a missing exponential in there (10^50?), since lead is an order of magnitude more dense than 1050 kg/m^3

Kind of nitpicking, but with an error like this I have to wonder about how well any of this is actually being explained. (Actually, given the accuracy rate of news articles on a science topic I'm very familiar with, I think I know the answer, unfortunately).

Reminds me a bit of Advogato.

BTW, personal commentary: I was instantly turned off by the front page. The whole 'superdev' thing instantly made me think 'ninja', 'guru', 'superstar', I threw up in my mouth a little and closed the tab without looking any further.

Don't forget ccache - storing the cache on a fast disk, it can easily speed up a build by 10x.

_Look at the success of phpBB and Wordpress_

Look at the number of horrific security holes in phpBB and Wordpress. Every other release or so fixes some new flaw.

> If he gave information about locations of jews to the Nazis, would we say it was freedom of information?

While we've got a time machine handy: If he gave information about the location and plans of Vichy officials to the French resistance, would we say it was an act of espionage/treason/terrorism and/or murder?

True, but there is a big difference - the 1040EZ is certainly easier to fill out, but the current convoluted tax laws mean that you could be leaving a substantial amount of money with the government unnecessarily. When I was first starting out I used the 1040EZ because I was scared of the complex forms - and later when I started itemizing I realized I had been overcharged easily a $1000 a year in taxes that I didn't need to owe had I simply filed my taxes differently (albeit using rather more complex forms).

It would be a very different story if the form suggested above was _the_ form, with zero loopholes available to people with expensive tax attorneys.

One difficulty here, if the idea is to identify policies that should be adopted country-wide, is confounding factors. Say you ban cash, and also introduce a 1984-style police state, and lastly pay for education and health care for all citizens. Crime drops versus the rest of the country. Is it because you're living in a police state, or is it because people are well educated and healthy? This difficulty seems especially problematic when some top level entity is a priori choosing what the new policies will be, versus choosing them as reality indicates.

A better system might be to have a number of startup countries, let the people in them choose what laws they will or will not have, and also allow people to move easily between countries, so if the policies enacted in one turn out to suck, people can vote with their feet. And with a large number of countries to choose from, the ones with the best policies will tend to win out over the long term. Say to start, let's have fifty of them.

No, the Washington Post is the sane DC newspaper. The Washington Times is owned by prophet/messiah Sun Myung Moon.

Actually, he seems to have a firm position:

"Innovation has been impeded in recent years by a patent system that too often grants low-quality patents with overly broad claims, which have been used by opportunists to extort royalty fees from manufacturers — particularly in the high-tech sector. The problem of low-quality patents is exacerbated by a litigation system that yields unpredictable and often overcompensating damages determinations, which divert investment and resources from innovation."

"This new legislation contains a combination of provisions that will improve patent quality; will provide a more efficient mechanism for administratively challenging the validity of patents; and will provide more discipline and certainty in damages calculations and willfulness determinations. The patent system relies on high-quality patents. Patents with overly broad claims or inventions that are obvious and not truly novel impede innovation and unfairly cast doubt on the validity of those that are high-quality."

From http://www.politico.com/news/stories/0310/34722.html

From Usenix 2010.

Abstract: Desktop computers are often compromised by the inter action of untrusted data and buggy software. To address this problem, we present Apiary, a system that transparently contains application faults while retaining the usage metaphors of a traditional desktop environment. Apiary accomplishes this with three key mechanisms. It isolates applications in containers that integrate in a controlled manner at the display and file system. It introduces ephemeral containers that are quickly instantiated for single application execution, to prevent any exploit that occurs from persisting and to protect user privacy. It introduces the Virtual Layered File System to make instantiating containers fast and space efficient, and to make managing many containers no more complex than a single traditional desktop. We have implemented Apiary on Linux without any application or operating system kernel changes. Our results with real applications, known exploits, and a 24-person user study show that Apiary has modest performance overhead, is effective in limiting the damage from real vulnerabilities, and is as easy for users to use as a traditional desktop.

You can buy credit default swaps in T-bills, so if the US defaulted you could make a nice return - if you think you can find a counterparty in the form of a large insurance company or bank who would actually survive a US default, which seems unlikely. An article in the Economist (IIRC) made the comment that these CDS's had best be denominated in Kalashnikovs and cans of spam.

Treasury rates are almost absurdly low right now, so we can clearly agree that there is very little perception of risk that the US will default. That's why it makes a great example of this phenomenon; people will look back and see all the (in retrospect) obvious signs of a coming default and wonder why it wasn't obvious now.

I would hope the structural differences between the US situation in 1945 and now would be obvious - exporter vs importer, young vs old, short term debt (the war) vs long term (SS and Medicare, plus maintaining an ongoing worldwide military presence). And we don't currently have the advantage that every possible economic competitor of ours was recently carpet bombed and utterly drained by years of total war, which certainly didn't hurt our exports.

Adding for clarity: I don't currently believe either a default or major coastal flooding is inevitable or unstoppable, I just think there is substantially more risk associated with it than is generally perceived. Ironically, the perception of little risk can help make such events more likely, because after all there is little point wasting lots of effort trying to prevent something that probably won't happen, right?

I'm confused by the 'likely to create an event in the next 30 years' part - either it's unlikely or it's likely. There are events that are quite likely, but with a low _perception_ of risk. Oil spills happen all the time, it's just been a while since one has happened in a rich Western country. Why? Because we've already tapped almost all of our oil - most of the worst spills happened long before we were born! Now the spills happen in places like the Middle East and Nigeria, but because only poor people live there, nobody notices or cares.

Other events that were or will be likely to occur, but have a low perception of risk, include the recent financial crisis, and, looking forward a decade or two, ocean level rise flooding major cities, and a US debt default.

Some interesting events that (seem to) have a genuinely low probability, but would really change the world if they happened:

* Asteroid hit (either a big one, or a smaller one that hits just right).

* A 'supervolcano' eruption that puts us into nuclear winter

* Solar flares fry every piece of electronics on the planet

* Sudden tipping point in the Antarctic or Greenland ice sheets that raises ocean levels by a few meters over a year.

* A strong 'godlike' AI (benevolent or otherwise)

* A self-replicating nanoassembler

* Godzilla attack

Probably obvious, but a lot of Aphex Twin and Orbital. I especially recommend Aphex's 'Surfing on Sine Waves' (done under the nym Polygon Window).

There is a guy who goes by Lichens who does some beautiful stuff - here is one from him http://www.youtube.com/watch?v=DfMjJ2oMXI8

Not strictly drum&bass or techno, but I'll additionally recommend: Explosions in the Sky (everything from them, but especially 'Those Who Tell The Truth Shall Die, Those Who Tell The Truth Shall Live Forever'), Godspeed You Black Emperor ('Lift Your Skinny Fists Like Antennas to Heaven' and 'F#A#infinity') and My Morning Jacket (especially Z).

All of the machine descriptions are written in Lisp. See gcc/config/i386.md for an example.

> At first blush, I dismissed it for violating 3.3.1

Even if it it did violate the clause: Apple changed it once, they can change it again...