Cellphone microwaves are good for brains?
Really weird. Though, it goes to show that argument that 'cellphones are certainly harmless because them have no effect because radiation is not ionizing' was probably BS. TBH this study does also reek of BS, given that they did not measure health effects of different levels of microwave radiation. I.e. they weren't actually doing much science, more like, cellphone advocacy.
By the way, the fact that mobile phone radiation is not ionizing (and is, in fact, VERY far from ionizing) goes only to explain why you don't quickly keel over and die after a single phone call (which you would had cellphones been using ionizing radiation, like x-rays or gamma rays, at same power levels).
That being said, I'll keep using mobile phone; epidemiological studies have not shown any health effects whatsoever, even though in the Europe virtually everyone is using mobile phone, and most people use it a lot more than landline phone.

The Polynomial: now in (stereographic) 3D

New update
Main highlights: 3D stereographic support: anaglyph and interlaced&checkerboard (some 3D hdtv's , Zalman TRIMON, etc).


The anaglyph mode requires cheap red/cyan glasses. You may want to use 'desaturate' slider in Visuals panel to make it work better for strongly coloured arenas.

The 'H stripes' mode put left and right images into even/odd rows of pixels. This works with some 3D displays such as Zalman TRIMON and various 3D HDTVs. It could also work with some shutter glasses system which use interlaced video signal (when each frame is sent as odd rows then even rows, and glasses flip so that those go to different eyes).
'V stripes' is analogous but uses vertical stripes. It could work with such thing and various commercial 3d displays that use lenticular lenses, as well as polarizing displays like zalman's but using vertical stripes.
'Checkerboard' uses checkerboard pattern, for this kind of thing (some 3D HDTVs do actually use checkerboard format).

You will need to tune the Span and Shift sliders for best effect, and maybe toggle the "swap sides" button if your glasses are other way around.

Todo: support for opengl stereo left and right framebuffer mode (probably works only with nvidia quadro cards though. I can't initialize opengl context in stereo mode on my pc at all, with geforce 9800 GT :/ ), and support for left image on one display and right image on other display.

Other good stuff: Changeable key bindings and various bugfixes.

Stuff that does not quite work yet but is coming: Multiplayer. For now, just ignore the 'devel' options tab.

microsoft got patent-trolled.

This is just insane. A Linux zealot I may be, but I feel for Microsoft in this case.
Patent law was intended to prevent a company from duplicating a better mousetrap without anyhow paying for the invention of that mousetrap. Patent law intends to prevent companies from ripping off the inventor. It is fairly certain in this case that Microsoft has paid engineers to develop a solution, which they did, without using anything whatsoever from i4i. Most likely before i4i showed up demanding the money, neither Microsoft nor their engineers have ever heard of i4i and their 'patent'. The patent covers storing text separately from its layout data. This is so trivial that any qualified software engineer would come up with it quicker than it'd take for him to read patent itself.
Programmers and software engineers are specifically selected&trained to be able to immediately 'invent' simple things on demand. Someone who can't immediately 'invent' this on spot wouldn't be working at Microsoft as software engineer.

Patent law is here to compensate inventors for their ingenuity - for creation of nontrivial things that wouldn't have existed for a while if not for the inventor. Patent law was never intended to let the 'first' expert who got a specific problem and straightforwardly solved it forbid others from independently solving this problem in the most natural and straightforward way imaginable.

The issue is that even the simplest solution could be presented as to appear very profound and mysterious to the totally clueless - just throw in a couple meaningless diagrams and other idiocies, and you get very nontrivial looking nonsense (see that patent, hiding triviality behind verbosity).
It is not understood by a court just how straightforward and retarded the patent is. The increased specialization results in perfect ignorance. It's perfectly equivalent to having 3 years old children for judge and jury - 3 years olds aren't stupid, but they do not know anything, they are not qualified to make any decisions on such case - and neither are adults whom do not understand the topic any better. Back in the day when patent law was introduced, the specialization was not so extreme as to prevent sensible trials; judge and jury could see if a patent is covering something totally straightforward.

On topic of Linux immunity

It's actually quite easy to write malware for Linux, no harder than for Windows anyway (and 'not running as root' wont save you). Fortunately, whatever the reason, very few linux viruses get written in first place; maybe the would be virus writers productively contribute to open source software instead.

More on antiviruses.

Just look at this. They claim there's 15,000 new virus definitions each day. For windows. How comes there's less than 1 new virus per day on other systems, which have only maybe 10x smaller marketshare? How come open source antivirus software has orders of magnitude fewer new definitions per day yet is fairly effective? The world is not so huge for such a number of new viruses a day anyway. How the hell are they counting, what are they counting as distinct viruses? Server side polymorphics? If a clever virus makes new variation every time, specific definitions aren't useful, you need a clever solution that lets the virus run but catches it when it tries to replicate.
My best guess is that this number is not even a count, it's simply a number that they figure is the optimal number to display in the software and write in their releases for the purpose of maximization of their profits and for advertising of their brand new "whitelisting" approach which ain't gonna protect anyone but would probably generate a lot of profit for antivirus companies (say, I launch a new software product, how it's going to get whitelisted if people aren't running it because its not whitelisted? The typical solution to a chicken and egg problem is that you have to buy a chicken. Or an egg, plus incubator. Meaning if there's whitelisting, developers have to pay for certifications).

In my opinion, antivirus is a broken solution to a wrong problem. If you run non-trustworthy code - such as pirated software, keygens for pirated software, various "toolbars", etc. or if you run email attachment, no antivirus can protect your (windows) pc - it'll eventually be infected. If you have insecure network services, antivirus won't protect you - but a security update to a service could. If you keep everything up to date and you don't run non-trustworthy code, then you're as safe without antivirus. A general security tool which watches for changes in files could be quite useful. A specific blacklist is of little use - it cannot protect even from variations of old viruses. A whitelist is just a nuisance. Antivirus software is written in such a way as to maximize profits of antivirus companies, not as to minimize threats; virus signature lists are far superior for profit generation than general solutions; it is far better to autorun files from usb sticks and then sell antivirus software than to forbid autorun for writable or all media. Microsoft's response of tightening OS security is the only hope for Windows world.

RIP Amarok 1.4

Amarok 1.4.10 no longer works for me under mandriva (installed from old repository, into /opt/kde3 folder)... it simply doesn't scan collection.

Meanwhile, Amarok 2 is still completely unusable, even though you can get rid of the "context". It's just insane. Can't resize playlist to take up all the remaining space, cant drag-resize columns in lists, and so on (resizing columns with a slider, who the hell came up with this?! That's outright insane. Anyone with a sense would fix the list control to get drag-resizing of columns to work again rather than make slider dialog for that).
I'm switching to Exaile for final, i think. I promised to fix some bugs in it, then left 'cause i had no time... gonna get on it

Are antivirus companies the main driving force behind virus writing?

I've always wondered why there is so many Windows viruses. Especially when I clean up friend's PC from malware. The number is on the order of hundreds thousands. An immense number of code lines. Awful lot of human effort. And when you think about it, the world is not so huge place. Naturally, some people suspect that some antivirus companies are somehow funding virus development, given that antivirus companies are the primary benefactors of virus development. This proposition, naturally, is commonly put down as an urban myth (no matter what).

I've figured some indirect but convincing evidence in favor of this 'conspiracy theory'. There's just far too many windows viruses and worms nowadays which replicate but do absolutely nothing besides slowing down the computer and saturating internet etc (so that antivirus speeds up the computer). Somehow, those viruses are the majority - viruses which actually do something like DDOSing a website, stealing credit card numbers, doing some evil as botnet, inserting obscenities into documents, and so on, are the minority - those worms are unusual, you read about them in the news. Even the botnets nowadays just sit doing nothing (Like conficker. A huge scare. It just penetrated into a lot of government facilities which it should not be able to penetrate into, which was quite seriously scary, and then did pretty much nothing except bringing billions into antivirus businesses).

This is very strange. That doesn't even look like vandalism or crime. Graffiti artists want their drawing to be seen; political vandals want to make damage to public property; criminals steal public property for scrap metal; all the IRL vandalism appears motivated, even if motivation is bizarre. There's always some driving force.
If you look at old dos (pre-windows) malware, nearly every virus did some original mischief - falling letters, animations and logos, inserted obscenities into the documents, wiped out hard drives, tried to say obscenities from PC speaker, messed with mouse cursor, and so on. Almost every 'harmless' virus did at least show a message about itself. There was some self expression, not unlike graffiti. You would expect most modern viruses to set something like goatse or 2girls1cup as desktop background, to scream from the speakers, to display political messages, to secretly record videos with webcam and upload those to youtube (particularly effective if combined with display of something nasty), and so on, a zillion possibilities. Indeed, that's what hackers do when they deface a popular website. But if you look at modern viruses, only a small fraction tries to do mischief or actually commit a crime. Majority seem to do nothing except supporting the antivirus manufacturers. There's almost no mischief and no graffiti. The viruses look like someone's boring daily job. Not like bored teens trolling. Okay, some nasty password stealers and such, those MAY be some criminal's daily boring job, but why harmless replicators don't even rickroll the user? (edit: actually there's a virus which rickrolls the users. It's on iphone!)

It seems to me that there is only one explanation: Development of windows viruses is nowadays heavily funded by antivirus companies - this at once explains why majority of viruses do nothing except replicating and generating scare, why amusing (when it's not your pc) virus pranks became rather uncommon, why there's very few Linux worms (mostly backdoors), and how it comes that antivirus companies 'detect' so many obscure viruses (which you would think user wouldn't notice) every day while being unable to respond promptly to real threats (which are extremely noticeable).

Antivirus company speakpeople would say that this is analogous to suspecting tire manufacturer of paying kids to knife the holes in tires. Well, firstly, that's an intentionally deceptive analogy. As matter of fact nobody's knifing tires in such a number as to sustain tire manufacturers; furthermore paying kids to knife the tires would've been far more dangerous and expensive, you can't outsource this to china or safely delegate it. That is why nobody suspects tire manufacturers, not blind trust that a big company would never commit a crime. They're making their profits by natural tire wear. Had they been making most of their profits from the tire slashing incidents, from unmotivated malice, then they would, in fact, be suspect (as the primary benefactors from the crime). The antivirus industry is more similar to heir inheriting billions from the rich uncle, who was killed by a car in hit-and-run near his house. Make that killed by a sniper shot - supposedly unmotivated sniper shot.

Secondly, as matter of fact, a lot of antivirus software is recognized to be fake - and the big brand antiviruses use pretty much same unethical tactics (popups telling you to upgrade, scaring you with numbers like '27 threats detected', reporting stuff like browser cookie files as threats, and so on) to generate revenue.

On the topic of trustworthiness of 'good guys'...
Putting aside small brand scareware, even the major 'antivirus' companies such as McAfee and Norton Antivirus engage in nearly fraudulent overcharging of credit cards of their customers (not outright illegal, but extremely close). If you did un-subscribe from Mc A Fee, they reportedly keep charging you the fee for 3 more months.
I certainly wouldn't trust such companies so much as to hold them above suspicion of virus development. There's certainly a plenty of ways to do this quite safely; e.g. a company could outsource virus identification to a separate company in a third world country, and this company in turn could hire a sweatshop of people and give 'em instructions vague enough that they could write the viruses in first place. Should this get discovered, the proxy gets blamed and liquidated, the sweatshop stays in place and keeps working (under different name). People whom were getting suckered into paying for antivirus still are getting suckered into paying for antivirus. People with a clue are 'outraged' but they would never have bought antivirus in first place.

I myself (I'm a Linux user) would not care about windows viruses and associated scareware at all if not for impact on the honest software developers. False positive rates of antivirus software are very high - the primary reason, i suppose, is that high false positives rate leads to increase in profits for antivirus companies - typical user tend to think that antivirus which found a virus is superior to antivirus which didn't find a virus. It appears as if some random short strings - which have nothing to do with any virus functionality itself and which appear in random software as much as in viruses - are consistently recognized as 'viruses' by design, resulting in credible virus scare for the customer. This is quite annoying for developers.

Ted Chiang's What's Expected Of Us

I strongly recommend to read this brilliant story (if you haven't read it before).

Some minor commentary. Consider the Free Will Device, put next to the predictor. Free Will Device is actually entirely deterministic, and doesn't have any free will of its own. It consist of photocell which watches the LED on predictor, timer, which gets reset to 0 every time light hits photocell, and actuator which pushes the button when timer reaches 2 seconds. If predictor blinks within those 2 seconds, there won't be a button press, and if predictor doesn't blink, there will be a button press. That's fairly deterministic and if you had been given a box and told that it works as Predictor in this story, you're bound to try doing exactly this - setting your mind to press the button if LED was dark for several seconds, to check if it really works.
The laws of universe in this story would forbid you to press Predictor button 2 seconds after the start of experiment or flash of LED (rather than 1 second) but not forbid you to take egg out after 3 or 5 minutes of cooking depending on your decision whenever you want it hard boiled. That's not mere determinism; that would require some special malice on part of the universe, forbidding you to make even a deterministic, predictable decision.

My view on 'crowdsourcing'.

Crowdsourcing is IMO much similar to common scam. Fraudsters are crowdsourcing they money. They make a business proposition which almost nobody would accept - to a large group of people. Very small percentage of whom make a decision mistake - i.e. get conned.
Same goes for most crowdsourcing. A crowdsourcer is making a business proposition - typically to write some software or make some design with non-guaranteed AND small pay - which almost nobody would accept (not even mythical people from 'third world'. Don't forget that both computer and internet connection are more expensive in developing countries). But with 'crowd' of hundreds thousands passing by, it is guaranteed that a few make decision mistake and accept. Bottom line is, both scammers and 'crowdsourcers' are profiting from rare psychological conditions and decision mistakes in a huge group of people.

Some crowdsourcing-like businesses could be different however. Innocentive, for example, where its mostly industrial chemistry, not programming, and rewards are quite big, good for few weeks work at $100+ per hour. Some of those problems might be nice for industrial chemist with relevant obscure expertise who can solve it quickly and win with no competition. I would call that expert sourcing; the industrial chemistry problems are of different league than programming and software design entirely and the spec work approach makes lot of sense when you really don't know if the problem is even solvable - and makes no sense what so ever if its mere matter of spending time. (I'm no chemist though sometimes I wish I were doing chemistry or physics for living)

Chicken feather hydrogen storage.

If this actually works, it is totally freaking awesome. Many fancy carbon structures can be formed by charring various organic wastes... even those that we don't know how to manufacture, e.g. fractal carbon structures.
Someone should try various bio materials for supercapacitors, in particular different kinds of charcoal.

Horrible Hashes

I've noticed something exceptionally stupid:
djb2 hash function.
http://www.cse.yorku.ca/~oz/hash.html
unsigned long
hash(unsigned char *str)
{
unsigned long hash = 5381;
int c;

while (c = *str++)
hash = ((hash << 5) + hash) + c; /* hash * 33 + c */

return hash;
}
and more popular version which has
hash = ((hash << 5) + hash) ^ c;
(The latter one is more prevalent. There was a topcoder match about finding collisions in latter one.)

The function itself is not all that stupid.

What's stupid is that if you search for djb2 on google, you see all sorts of people 'personally recommending' it as best and fastest simple hash, people trying to explain why it is good (when the answer is: it is NOT good), people wondering why 5381 is better (it's not), people tracking the history of this "excellent" function, naive people whom trust this sort of advice using it in various software (ruby's hash tables?), etc. All in all people presuming that 5381 and 33 got some special significance and are much better than e.g. 0 and 31.

What is so stupid about it? For starters, even though the output of this function is 32 bits, not even for the 2 char ASCII strings do you have a guarantee for lack of collisions. In fact "SV" collides with "Pt", and "g5" collides with "as" [in the second version that uses xor], just to name two examples out of hundreds. Each character except first provides only about 5 bits because that's how much you get out with *33. That's not good. That's complete crap. From a 32-bit hash, you can expect to get no collisions at all between 2 character strings, especially restricted to lowercase or uppercase alphanumeric. Most primes work no worse; you can use 257 and then your function at least will not collide on 2-character strings (it will still be crap though, especially if you use parts of hash; this doesn't need to be a prime, only needs to be odd and you ought to run code to select best for hashing some real data like list of all file names if you want a good number, not look on the internet and judge by people's reputation). Furthermore, there are a lot of collisions between strings that differ by 2 characters, because 2 consecutive characters can be altered to keep same 'hash'. Got to give some credit though. In some very limited original usage (hash table of specific size, with specific key statistics, e.g. English words), which I do not know, and which you are highly unlikely to replicate, it may have been good. Or actually, not good. Merely not too bad.

What is the significance of 5381 ? Apart from low 8 bits of 5381*33 (in the variation which has xor instead of add), it is pretty much totally irrelevant to collision resistance, it is just multiplied by 33ⁿ and added in. This function is pretty much as crap with start value of 5381 as with start value of 42 or 100 or 12345 - the only difference is that unexplained 5381 hints at some deep wisdom whereas 12345 does not.
Now the structure. The new version of that hash function has multiply and xor. This function sucks as much with xor as it does with addition, but with xor it looks better and harder to analyze than with addition.

Moral of the story:
Do not trust magical looking code. 99.9% of magical looking code out there is utter shit, only obfuscated enough that you can't immediately recognize that it is shit, and most of the time 'smart' looking constants are also really shitty choices or are at best random choices (at the VERY best, magical constants could have been selected for some very particular case which you know nothing about, by a method which you know nothing about, and are still most likely than not bad for whatever you want to do).
Do not trust internet advice or consensus either. Keep in mind that majority of acclaimed programming experts are expert only on one thing: persistent bullshitting to obtain 'expert' status and sell some books - they're totally mediocre programmers and software engineers (doesn't apply in this case 'cause qmail seems kind of good, but nonetheless).
Especially the old usenet celebrities. Not even if they have a wiki page about themselves, not even those whom somehow got PhD. Also, keep in mind that majority of people in 'consensus' are simply repeating each other, like parrots, and haven't devoted much brain time to thinking about the question.

This is why science does not and cannot function by reference to authority, but only by reference to argument, to actual reasons, and why if no reasons are given you shouldn't assume that any exist.

edit: also, don't even get me started on "fast". If you want fast, you'd better do 4 chars at once, on 32-bit machine.

edit: clarified on the version with + and version with ^, even though those have very similar properties.

What's with all those design concepts?!

What's with all the infestation of internet with "green" design concepts that cannot work but win awards?
I mean designs such as
The amazing plaane of the future . (With a freaking wind turbine on the tail! You use wind turbine to power the engines with the energy of motion of airplane through air!)
Gravity lamp (debunked)
mp3 player powered by spinning finger in a hole.

The mp3 player one has not been debunked properly yet, but it is very similar to this lamp in that it utilizes common lack of intuitive relation of mechanical to electrical to sonic or light power.
Assuming hundred percent efficiency, headphones consume 20 milliwatts at max power, or 1.2 Joules per minute. 1 Joule is about the energy of 1kg lifted to height 10 centimeters. Spinning a finger in a hole is about the least ergonomic way to generate power; the smallest possible leverage in the least convenient way (and you can't spin player around the finger 'cause of headphone cable). Lifting 1kg to weight 10cm every minute by spinning finger in the hole is obviously out of question, unless you're doing it constantly. (Furthermore, the efficiency of mp3 player is far below 100% due to the power consumed by mp3 decoder).

All in all, a hand powered mp3 player, lamp, or other 'low power' appliance will need occasional but fairly vigorous spinning of a crank, squeezing, or vigorous shaking (e.g. during exercise). It would of course be very cool to power mp3 player by occasional turn of a finger, absolutely amazing in fact because such player would have to include a perpetual motion device, that's why this sort of stuff seems cool and amazing.

And while you're at it, forget about powering laptop by opening a lid or with power of keypresses - it is possible but such laptop a: won't have backlight, b: won't run any modern applications (think of having 1..10 mhz cpu , with the computational power of pc from 20 years ago), and c: but it would work for months on regular battery and could recharge by solar, rendering the whole keystroke power issue nill because solar panel is going to take less space.

On topic of energy saving measures, I limit ThePolynomial's fps to your display's refresh rate by default; this OpenGL feature doesn't seem to be supported on ATI under Windows (according to user reports, didn't test myself), which if true is really downright despicable behavior on AMD's part (they probably do it because of the few gamers whom would think ATI having 1000 fps and NVidia having 70 fps in games that syncs to refresh by default would make ATI look better, or some other silly marketing related reason for not implementing refresh syncing).

Multiplayer progress.

I've been working on multiplayer for my game for past few weeks, hence the gap in updates. Turns out good multiplayer is kind of difficult, and I'm somewhat behind my schedule, but expect updates soon.

TopCoder: lying again.

background: TopCoder is a programing spec work business (spec work is called "crowdsourcing" nowadays). They also run some regular programming competitions (which are not work for hire), sometimes with okay problems, sometimes with so-so problems, sometimes with problems that 1/3 of participants can solve exactly. I competed there a little once in a while purely for self evaluation purposes - they do somehow have a big community, and there are a few good programmers on top to compete with.

Anyways, where was I... yes, TopCoder lying in their press releases.
It's interesting how a company can't learn a lesson that lying in public releases is not always a great idea. A while ago, they had hired some girl in china - she may have been a good choice for the job - I've no information about this - and then lied a shitton about her qualifications and achievements [see original TC's press release which was then echoed by girl's university] resulting in massive PR success followed by even more massive PR fail in the china, totally ruining girl's reputation. The lying, for a public release, was not very outstanding - just massive exaggerations, pretty standard for small company's public release, a small company has to look big, but it did ruin the girl's reputation 'cause of cultural misunderstanding, its not everywhere customary for a company to exaggerate how great their new hire is. On darker side, I bet they got her to sign their "affidavit" beforehand which explicitly forbids you from suing TopCoder for damages arising from this sort of misrepresentation of you. [you need to sign this at notary if you participate in competition and get a prize; that's quite serious. I won a prize at TopCoder once and asked for legal advice on their affidavit, a friend told me of that girl's story, which I remembered 'cause its really scary how individual could get chewed up by gears of commerce and spit out]

Recently, there had been a "NASA-TopCoder" contest with '25 000 $ in prizes'. It seemed a little strange.

The NASA-TopCoder Challenge will be the first time the TopCoder community of more than 220,000 software enthusiasts is utilized by the world's leading aerospace organization. Long-term human space missions such as those being planned for Mars, will require higher levels of pre-planning and more analysis of available data than ever before. Biometric modeling and simulation programs are algorithmically-intensive as flight surgeons explore and evaluate every possible medical scenario that might occur on long-term missions. In this experiment, competitors will develop algorithms to help NASA's flight surgeons make decisions involved with optimizing the contents of the medical supplies kit that may one day be carried onboard long-term space missions. The submissions will be compared with the results of an existing computer model that has simulated the expected medical occurrences and outcomes for various mission scenarios.

Under closer examination (I registered for the contest because I was rather curious and because invitation email didn't quite made it clear who funded the experiment), it turned out that it indeed was a business research experiment (25000$ from research grant from some business university were used to run 24 tiny contests in parallel for some sort of business research). Needless to say, there were no NASA representatives on contest forum answering the questions about problem or asking questions about solutions [correct me if any did show up since I lost the interest]. Nothing of this sort. Typical programming competition, with a typical competition problem that has only superficial resemblance to real requirements for real software. Very simple model - much simpler than your 'model' when you visit pharmacy and decide what to buy. In real life if you get a splinter under your skin, you will need tweezers to remove it. Then you can use hydrogen peroxide or you can use iodine, or other antiseptic, and if you don't treat the cut with antiseptic you might need to use topical antibiotic later to treat inflammation. That is not simulated in contest - the supplies are not ever interchangeable and medical conditions are not dependent on prior conditions and treatment (worse than that, them are totally statistically independent from prior conditions). It's absurd to think that contents of medical kit for a Mars mission would be based on such simplistic assumptions, so much more simplistic than the ones you'd make when you visit pharmacy. Yet participants would believe it because it's happier to believe you contribute something to space exploration.

Furthermore, interestingly enough, in the "community of more than 220,000 software enthusiasts", only about 1700 registered and only 400 participated in the contest.

As NASA source indicates, the truth boils down to this:

The competition originated when professor Karim R. Lakhani of Harvard Business School and professor Kevin Boudreau of London Business School invited NASA to provide a compelling technical challenge to monitor and analyze the results from an open innovation management perspective. Their research project is funded by grants from the London Business School M-Lab and the Harvard Business School.

, and naturally "topcoder asks for and gets a simple contest-style problem from NASA for use in their business experiment" is a whole bit less impressive than "NASA employs topcoder to solve something for human spaceflight".

Version 00h

Minor changes for more compatibility with various Linux distros, should work on ubuntu karmic now.

Cutting a disk from aluminium...

So, I need to cut a disk from aluminium plate...



Various parts: pieces from a wooden box, motor and gears from canon i250 printer which I picked up for parts other day, outdoors plastic chair, transformer from very old tv for 7v power (printer's powersupply was missing), a paperclip, piece of file, drill bit.

It features a very advanced and hi tech design, with automatic pressure adjustment (when this thing jams, pressure on cutter decreases 'cause of pull on that rope). It's going to take a long while to cut through tho, motor isn't very powerful, but as long as I'm not spinning this by hand, I don't mind.

Side idea: with 3 old printers, you can make a neat CNC machine or even reprap machine. You'll only need to cut a thread onto the paper feeding rod, and use it to move the table. There's even a very precise position encoder here. If I find 2 more broken printers, I might make this.