I pushed out a new BitTorrent release, but it's seriously borked because of sourceforge. Thanks guys. This is the third release in a row which has gotten garbled - only the last two were nicely garbled, this one is flat-out giving 404 most of the time.
Codecon is coming up this weekend. It's only $95 for three whole days of presentations, everybody should come!
Codeville
Speaking of which, Codeville, a project by me and my brother, is much more mature than the last time I mentioned it (check the page for details) and will be presented on friday.
The program for CodeCon 2004 has been announced.
http://www.codecon.org/2004/program.html
CodeCon is the premier showcase of active hacker projects. It is a workshop for developers of real-world applications with working code and active development projects. All presentations will given by one of the active developers, and accompanied by a functional demo.
Highlights of CodeCon 2004 include:
The third annual CodeCon takes place February 20 - 22, noon - 6pm, at Club NV (525 Howard Street) in San Francisco. CodeCon registration is $95; a $20 discount is available for attendees who register online prior to February 1, 2004.
Just a few days left to submit to Codecon 2004. It only takes a few minutes, get those submissions in!
Is ZF set theory a hack?
Raph links to some discussion whether ZF set theory is a hack. The up shot is that all the different proposed formulations for the foundations of mathematics are several times as large as you would expect.
I've long thought that the one true criterion for acceptability of a basis of mathematics was its intuitive acceptability to a human. Unfortunately there is a long history of humans, including professional logicians, forming bases which later turned out to be inconsistent. Now we know that our intuitions about what is simple are also completely wrong.
Perhaps instead of judging based on untuitive acceptability, a fundamentally subjective and poorly defined criterion, we should instead judge based on number of symbols in representation. That at least is a well-defined and measurable concept. Maybe its possible to make a basis which is based on much more abstract concepts even than set, which requires some work to build even our most basic intuitive concepts, but is much simpler at core.
Formulating a better foundation of mathematics may also help with computer proving. Perhaps we are crippling our theorem provers by forcing them to view mathematics through the lens of human intuition due to our selection of an extremely cumbersome set of base axioms.
Clickmazes
clickmazes.com is cool.
Twisty Puzzles
I came up with some interesting twisty puzzle ideas.
Smith Numbers
Smith numbers raise some interesting philosophical questions.
Jumping Champions
Jumping champions are neat.
Ebay's system of minimum bid increments can be gamed slightly. Let us say that I'm selling an item which is currently at $7000, at which point the minimum bid increment is $100. I could now have a shill bid $7100 and have full confidence that this would definitely make the price go up, since if the highest bidder bid less than that the price would have been at their highest bid.
Shill bidding is against ebay's rules, and there is active policing to stop it. However, detecting gaming of this sort done on a small scale is very difficult. In fact, I might not even have known about it - maybe a friend of mine just bid the $7100 to help me make some more money. A technical solution would be much better than a policing solution.
As it turns out, there is a simple technical solution. In the (rare) case where the highest bid is less than a minimum bid increment greater than the second highest bid, the current winning price should be listed as the second bid plus the minimum bid increment, even though if the current high bidder wins they'll get it for their high bid. If someone else bids higher the winning amount should change to what was previously the second-highest bid plus two times the minimum bid increment. Then shills couldn't outbid without running the risk of getting the winning bid and blowing the sale.
Surprisingly, this system doesn't result in the winning amount being dependant on the order in which bids are placed. The winning price is the greater of the second highest bid plus the minimum bid increment or the third highest bid plus two times the minimum bid increment. Or the maximum the winner bid, if it's less than that value.
So this rule causes some extra complexity in calculation of the winning bid, but not very much. It also adds some potential confusion to the interface, but that can be minimized by making the high bidder see their real winning value while showing the fake (and only slightly different) one for everyone else. Whether these problems offset the benefits of stopping such a simple and likely widespread form of shilling is ebay's decision, but I hope they at least read this and consider it seriously.
There are of course much simpler (and potentially lucrative) forms of shilling. For example, one can have a shill bid which is quite high, then later send mail to the real highest bidder claiming that the high bid was a no-show and offering to sell it for their maximum bid. If you get mail like this, don't immediately assume that it's a cheating seller, since bidders are occasionally no-shows. Instead, you should respond by offering to sell it for what the price would have been if the high bid had never happened. I'm about to do that right now.
I have to say I'm impressed with how much improvement ebay has made over the years. I was going to post about some much more serious errors in their bidding system, but it turns out those have been fixed since the last time I studied it.
The Codecon 2004 Call For Papers is now out. If you've got an actively maintained cool project, please consider submitting.
Bram's Law
ncm: Sadly, artificially causing difficulty is an invalid loophole. But adding functionality which is inherently more difficult may work. Bloating up what should be a simple spec is mostly just an example of Bram's law in action.
Databases are a good example. MySQL started without transactions at all, has no indexing for joins, and doesn't scale to a decent number of rows even with that ridiculously limited feature set. Postgres, on the other hand, supports full transactions, lots of fancy joins, and scales up extremely well. Unfortunately, Postgres can't be simply pointed at a file and used as a library, which is why people use MySQL.
Image Grouping
I had an interesting idea for a user interface for image grouping. The user is shown three images and asked which one least belongs. Then another three, and another three, etc. I'm not sure what the point of this is, or what one might do with the data, but it seems like an interesting idea.
What Customers Want
The things which will make people love your software, by rapidly plummeting order of importance, are:
The order of priority many people use when writing software, and, unfortunately, what users generally say they want when asked, are:
This is a siginificant discrepancy.
The last trust metric I posted about can be improved significantly both in terms of run time and behavior by switching from number of certs in to number of nodes certed.
Make each position be a float, rather than an int. At each round, lower each node by one unit for each spam it sent out. Then, for each node which is below one which certed it, raise it by ten units and lower lower each of the ones which certed it and are above it by ten units divided by the number of them.
I think further dramatic run time reductions are possible.
Lightning Thermal Energy
The transfer of energy to the outside of a lightning thermal plant doesn't have to be done with anything as fancy as microwaves. A simple pair of tubes, one with air going out, the other with air going in, leaves the energy in completely mechanical form and transfers at high efficiency without acting as a conductor.
ncm: The largest solar plants are actually solar thermal instead of using photovoltaics. This would seem to indicate that solar thermal is just plain cheaper. As for doing anything with lightning, lightning doesn't follow the normal rules you're taught about in electronics class. Any change which can jump through the atmosphere isn't going to much care about a few piddly feet of vacuum trying to act as an insulator. Attempts to change batteries with lightning have gotten a few people blown up, and attempts to charge capacitors with them have mostly resulted in burnt electronics. With regards to your specific idea, if you acatually got the electronics to work (doubtful) it would probably result in the object you were trying to lift being turned into powder or, worse, shrapnel.
Here's my plans for a lightning thermal energy plant -
Go to a place which gets lots of lightning. Dig a deep hole. Extend outwards from the bottom of the hole a bunch of spokes of some highly conductive substance. At the bottom of the hole put a conductive plate attached to all the spokes. Fill the hole with carbon. (Or maybe with tungsten. Carbon is cheaper but more fragile.) On top of the filled hole put a ventilated faraday cage with low heat conductivity. In the faraday cage put a stirling engine. To the top of the faraday cage attach a very long lightning rod.
The idea here is to get the lightning to go through the carbon despite it being a resistor. There are several technical problems here -
Anyone with more engineering knowledge than me know if these problems can be overcome?
A bunch of people informed me that the solar enery harvesting technique I came up with is an example of what's called 'solar thermal energy'. Zaitcev has a good analysis of my exact proposal. (The upshot is, it's a lot easier to get the water hot enough to boil it then use a plain old rankine engine. Parabolic dish ones tend to use stirling engines.)
The biggest current solar thermal energy plant is a solar tower. A cheaper design (when done at sufficient scale) is a solar chimney Solar chimneys are a monument to inefficiency but in the desert we have more solar energy than we know what to do with, so the real issue is cost rather than efficiency. The robustness and simplicity of solar chimneys is certainly appealing. After the cost of construction is amortized out after thirty years or so, it produces energy essentilly for free and with hardly any maintenance.
Strangely, the solar thermal plants don't use fancy new kalina cycle engines.
My next question is, could lightning thermal energy be viable?
In other engineering news, there's a new desalinization trick.
Twenty years after the author's death or the author's hundredth birthday, whichever comes last -- that's a workable standard to provide for the author and his or her immediate heirs.
Among the many obvious things wrong with this statement is a more subtle and alarming one. He's making the implicit assumption that money must come dribble in over time, as if it's impossible to save and invest. There is a very deep-seated belief in the consumer culture of the united states that any money one has must of necessity be pissed away, and the only way for old people to survive is to receive a pension which they can't borrow against when they're younger.
Those of you who understand savings may have trouble beliving anyone actually believes that. Let me assure you, most people do. The next time someone makes a statement which implies that savings is impossible call them on it. You'd expect them to say that obviously you misunderstood them, but more than likely they'll actually arague with you.
Solar Energy
Solar panels are expensive. I've come up with an idea for how to change solar energy into electricity which at least sounds cheap. There's probably something completely impractical about this scheme, but I'd like to hear what the practical difficulties might be from someone who knows more about mechanical engineering than I do.
There is a big tray of water with an airtight enclosure around it. During the day, sunlight heats the water inside. Air is pumped out of the enclosure until the water inside boils because of the low pressure. The resulting steam is sent through a pipe where it turns a turbine which is used to generate electricity, then through pipes in a water tank which is kept in shadow beneath a mirror to cool off. When the vapor condenses back into water it's funneled back into the tray. At night, the heat shielding on the coolling water tank is taken off so it can cool off again for the next day.
It's a big mostly passive device which generates electricity as long as the sun keeps running. It sounds cheap and efficient, but I don't know anything about the difficulties of making a vacuum chamber that big, pumping out that much air, or keeping the water from boiling directly back into the coolling tank. I suppose the boiling back could be fixed by making the process two stage; First the water boils off and is changed back into water in the coolling tank, then the coolled water is let back into the tray en masse to be heated again.
Update: Of course I realize right after posting that that you could set up a bunch of mirrors to make a really big solar oven so the water boils at normal atmospheric pressure rather than having to play pressure games to get it to boil.
New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.
Keep up with the latest Advogato features by reading the Advogato status blog.
If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!
