Man, is there even anyone on Slashdot stupid enough to fall for this shit?
Yeah, I can see the morons on Facebook and Youtube being technically incompetent and gullible enough to fall for it, but here? Half of us aren't even running an OS this "software" runs on - hell, I'm pretty sure at least one regular user here is running an OS he wrote in his free time.
Video game graphics could probably benefit from this. Very few people will notice that one pixel is #FA1003 instead of #FC1102, especially when it's replaced 16ms (or, worst-case, 33ms) later with yet another color. It might actually make things "better" - making the rendering seem more analog. Many games are "wasting" power adding film grain or bokeh depth-of-field or lens flares or vignette, to try to simulate the imperfections of analog systems to try to make their graphics less artificial-looking. If you can get a "better" look while using *less* power, all the better.
Actually, I seem to recall hearing about this earlier. For some reason I want to say nVidia specifically has been looking into this.
This is a little mini-speech I like to give to people, and it's rather appropriate for the this:
In the US, politicians train and study as politicians. They have degrees in political science, or law, or economics, or maybe history or business. Obama was a lawyer. So was Clinton. Bush II had an MBA. Bush I studied economics, as did Reagan. You have to go back all the way to Carter to find a president that had any sort of remotely "practical" training, as a naval officer specializing in nuclear submarines.
In China, politicians train and study as engineers. If shit goes down (as it is wont to do) and the revolution comes, President Hu Jintao could flee to the US, change his name, and live out his life working as an engineer (hydraulic engineering - his first real job was at a hydroelectric plant). Vice President Xi Jinping studied chemical engineering. Premier Wen Jiabao studied geomechanics. Wu Bangguo studied electrical engineering.
Engineering, fundamentally, is "the study of solving problems". It's not, strictly speaking, a science, but an application of science to the real world.
Modern American politics seems to be less about "solving people's problems" and more "making new problems to 'solve' so you can stay in power".
In case you haven't noticed, China is beating us. They're obviously doing something right, and I don't think it's the censorship or the market controls. Their system of government may not be better than ours on paper - slow, central control of everything rarely works for long - but they're better in practice because they have people who actually do the job they're supposed to do.
NOW is our chance. The Chinese seem to be making exactly the mistake we made - their up-and-coming leaders are career politicians, born-and-raised to rule. At the same time, their population boom will be shifting from a worker-heavy populace to a retiree-heavy populace, causing exactly the economic problems we're having now with all the Baby Boomers retiring.
We get one chance to get back on top. We need a government that responds to us, one that works quickly and efficiently for the benefit of everyone.
If anybody knows of any good candidates, speak up. I do not want a lawyer to represent me. I do not want a manager to represent me. I want an engineer, a man (or woman) who solves problems, because we have a lot of problems that need solving.
Does it? I think it's a valid comparison, because it's fundamentally the same sort of situation. Both "services" have both legitimate and illegitimate* uses. Most people would argue that shutting down the freeways would be blatantly wrong, as it harms the vast majority of legitimate users far more than it harms the minority of illegitimate users. So the question then becomes "at what point do you 'shut down' a service that has both uses"? What ratio of illegitimate to legitimate users is necessary? 70%? 50%? 20%?
The obvious "answer" is "when more people use it wrongly than use it rightly", but as with all easy, obvious answers, that is also demonstrably wrong. Take, for another ludicrous example, amphetamines. The illegitimate users vastly outnumber the legitimate, but you'll note that it is not completely banned. Heavily regulated? Yes. Illegal usage punished? Yes. But completely, 100% banned? No, because there are still proper medicinal uses for it. And you can find thousands of other examples, from leaded gasoline to automatic weapons.
So now we've established that a total ban on something with any amount of legitimate use is, at the very least, not an accepted practice. We don't need to rely on abstract philosophical arguments - we can point to concrete examples. So we've essentially "proven" that you should not ban Bittorrent, inasmuch as you can "prove" anything in as loose a field as ethics and law.
So now the question goes from "do we ban Bittorrent?" to "how do we stop illegitimate uses of Bittorrent?", which is the question we really ought to have started with. And that question I'm afraid is too complicated for me to continue delving into.
It's only a bad example if you don't think. Admittedly, getting John Q. Public to actually think may be difficult...
* I'll note also that even "illegitimate" uses of Bittorrent can be legitimate. I've gotten into the habit of torrenting certain games I own, simply because I don't want to be bothered putting the CD every time I want to play. Completely legal, in my case - the 4,999 other people in the swarm may or may not have a similar justification, and I'd probably bet on the "may not".
I have to say, some of their comparisons seem... unfair.
Their main chart compares three things: regular "console connected to display", "current cloud systems", and "GRID cloud rendering".
First off, they cite 66ms latency just at the display level, which is definitely at the higher end of the spectrum. But at least they use the same latency in each.
Their cloud/cloud comparisons are also quite suspicious. Reducing encode by 60%, yeah, I can buy that. Reducing *decode* - which, I remind you, is done client-side - by the same amount is also suspicious in light of their "this does not require an nVidia client, it will work with any h.264 decoder" claim.
Then they claim to have reduced network latency, and significantly (75ms to 30ms). Now, I can vaguely see how they *could* - if they can reduce bandwidth usage significantly, they might eke out a bit less latency, but I highly doubt they can more than *double* their compression efficiency. Unless they're doing something crazy like putting a network interface directly on the GPU (one image they show contradicts that theory), I think this claim is also pretty dubious.
The worst one is the "game pipeline" time. While I can believe that a newer, more powerful graphics card can definitely perform *twice* as well as an older one, I can also state definitively both that "you can put that new card in the home console as well" and "new games will expand to use that power, leaving you back where you started at 100ms render times". The former I can state because they've *already* released Kepler-based cards (to rave review, although my own seems to be backordered); the latter I can state because that's how the industry has worked since at least the late 70s.
Long story short, they seem to be doing some *extremely* unscientific, biased comparisons. Do they probably have something here? Yeah. Is it literally going to be as good as an actual console (or better yet, PC) connected directly to the console, as they claim? No.
Okay, so -gate has been genericized. Happens to almost every word given enough time.
First, it was used for one specific instance: Watergate. Then we had Koreagate, Billygate, Monicagate - any Federal-level political scandal.
Then it went international: Dunagate, Mabelgate, Petrogate - any political scandal
Then it went non-political: Closetgate, Climategate, Cablegate, Crashgate, and those are just the ones that give me an alliteration bonus. Oh, and Bonusgate.
Wikipedia counts one hundred and nineteen distinct -gates beyond the original Watergate (and also excluding fictional -gates). This includes two distinct Memogates, Nannygates, Grannygates, Spygates, Strippergates and three Troopergates. And those are just the ones meeting Wikipedia notability guidelines.
Languages *change*. The news media uses meaningless sound-bite catchphrases. History gets distorted. Get over it.
Seriously, what features? I've never tried to write anything as low-level as a kernel, but I can't think of much compiler-specific stuff besides a bunch of compiler detection to set up properly-sized typedefs. Maybe differences in assembly syntax?
Also, clang was designed to be a drop-in replacement for gcc in many (not all) ways. Might make it easier for this case.
I've found that code that will compile properly under a variety of compilers tends to be of better quality.
One of my current projects started out on an old 2.x branch of GCC. When I finally got around to updating to a current GCC, I had to fix quite a few bugs before it would actually work - the different compiler was catching problems I hadn't noticed before.
Same when I tried compiling it under Visual Studio, or Clang - the more compilers I made it work under, the less bugs there were in the code.
Now, if a given program actually uses some special feature of GCC, that's fine - if only one compiler will do what you need, that's fine. Or if it's too much work to maintain a "port" - I stopped maintaining the VS project files a while ago, since I no longer used it. But if you have a chance to at least test it against a different compiler, go ahead and give it a shot.
Fact: The standards, Unicode in particular, do not specify one or two lines in the "currency symbol". That is left to the font to decide.
I learned this while setting up a currency database. Apparently Brazil (I think it was Brazil) uses the double-barred symbol for *their* currency, and the single-barred for "US Dollars", which are also in relatively common use. Pretty decent idea for distinguishing currencies, IMO - is sure beats US$ vs CA$, or using ISO 4217 codes.
I'm probably a decent subject for a "do the damn kids even know what this stuff is anymore?" experiment - I graduate college in a few weeks. I know, I know, that's a hell of a selection bias right from the start, but let's go with it.
Simple test: go through my smartphone, see if I can ID every icon's source on one particular screen.
First, the phone itself has a handful of icons. There's an arrow pointing backwards for the general "back/undo" button, a series of lines (with the top extended to the left) for "details/menu", a house for "home page", and a magnifying glass for "search". The first two will never really fail, since they were abstract from the start. Houses still have walls, a floor and a roof, although chimneys may be inexplicable in a few generations. Magnifying glasses are still seen at least on TV, where they're heavily associated with "detectives", making "search" a logical connection. Finally, next to the power button are two logos: a vertical line enclosed in a circle, and a padlock. The former has always mystified me, but is widely-used to mean "power device on/off" (usually, but not always, the circle is broken at the top). The latter, however, is still common enough a sight.
Next icons are on the notifications bar. First one of those is "unread voicemails", which I can recognize as an old cassette tape, although I definitely associate that more with its icon use than its physical use, and I will concede that some people my age may never have used one. Then is a square with a triangle in it, indicating that the media player is playing, which is pretty much a purely abstract icon. Then are two different sets of bars for Wifi and mobile status, again purely abstract AFAIK.
Then there's a little image of the phone surrounding by motion lines, indicating "vibrate mode". Since it looks pretty much like the actual device, I can definitely recognize it. Then there's a battery life indicator, and yes, we do know what AAs are. Last notification icon's the "hardest" - an alarm clock, looking like the traditional round analog clocks with the bells on top. I no longer actually use a dedicated alarm clock - I use this phone instead, but the concept of "a distinct device used for keeping time and playing alarms as specified" is not alien to me. Although I never owned an alarm clock shaped like that, I can recognize the round analog "clock", and from there, even without seeing it a million times in pop culture, I could derive "alarm clock" from that.
The main screen has more icons. There's a plain, old-fashioned computer monitor for my terminal-emulator app, a piece of paper and pencil for my text editor app, a spiral-bound calendar for the calendar app, a set of arithmetic symbols on buttons for my calculator app, and so on. The only ones that someone my age may not recognize are: * the old-style red/blue 3D goggles for the "Google Goggles" app * the index-card thing for the Contacts app * the globe icon for the web browser - nobody calls it the "World Wide Web" anymore, so the connection between "world" and "internet" is now a rather loose one. * and fine, I'll go ahead and list the chess board and knights from my chess app, since I know half the people my age never learned to play chess
Just as a note, on my phone, at least, "Settings" uses a dial, not a set of gears or a wrench. This arguably makes far more sense - a dial is used to change a setting, while gears are what makes something actually work, and a wrench is more used for "assembly" than "configuration".
There's also a few icons that are *new* and still meaningless - can someone explain the Bluetooth logo?
Let's go into one app and see what icons are in *it*. For this experiment, lets try the media player app.
Well, there's a "series of bulleted items" that brings me to the current playlist. Logical. There's two different arrow-based icons for "random" and "repeat", which are also self-explanatory provided you know how abstract arrows work (aside: does anyone know if these abstract ar
This thing is cool and all, don't get me wrong, but the video was a bit of a letdown. Very quiet (when the blog states one goal was "Have lots of moving parts (preferably loud ones)", and not a very good show of how it actually works. It didn't even use 2/3rds of the "digits" - the video stops after reaching 8, so only the first digit of each register was used.
First, installing something on the suspect's computer *generally* can't be done stealthily. If the FBI knocked on my door, handed me a warrant, and installed spyware on my computer, I sure as hell won't be doing anything even slightly suspicious on my own computers - a quick hop to the library, or to a friend's house, and then use *theirs*.
They may also not know where the person is. Say they're trying to catch Steve McBadguy, who's "on the run". They know he tends to log in to his Facebook from public computers while he makes his way towards some country with no extradition laws. They could put a tap on every public computer between here and Uzbekistan*, or they could put a tap on one specific Facebook account. I believe I speak for everyone when I say the latter is better for *everyone* - more privacy for us, and less work for the FBI.
* Fact: The United States does not have an extradition treaty with Uzbekistan.
What happens if, say, you get a virus, or your app goes haywire, and The Critical Irreplaceable File gets corrupted or deleted?
RAID protects from hardware failures. It does nothing against software or human errors. And given my history with hardware, software and people, I'd say the first is generally the most reliable.
Look, if it's a data stream, you can record it. I'm not saying everyone should have an API that the FBI can use. I'm not saying we need to record absolutely everything so the FBI can look at it.
What I'm saying is that if the FBI needs to record something and they have enough evidence to get a warrant, they can come in and write their own damn code to log it, we'll put it on the server for as long as the court order says, and then as soon as they're gone we revert the code back to the way it was. Or, the FBI can log every packet themselves, and *they* get the fun task of sifting through billions of TCP packets to find the ones used by Ahmed ibn Badguy.
And if the system *is* anonymous-by-design, well, "that's literally impossible" is generally considered a valid reason to refuse a warrant. I know if the FBI knocked on my door and handed me a warrant for "whatever is 40km beneath the property" and a shovel, I'd call up the judge and tell him that, unfortunately, the laws of science trump even the US Constitution.
Where, exactly, did you get "private business bearing the burden of law enforcement" from what I wrote?
The FBI would either a) install a relatively simple network device themselves, requiring at most a few minutes downtime, b) write some basic logging code themselves, or c) compensate the private-enterprise-programmers for doing (b).
First, the FBI gets a warrant for a particular "wiretap". This should be absolutely mandatory for what I'm about to propose.
Then, off a specific warrant, they go to whichever company the warrant lists, and either:
a) Install a packet-sniffer in front of the web server, logging everything to disk, which is then physically taken by the FBI as evidence - just like a conventional phone wiretap. This avoids the whole "anyone could use the backdoor" - if "anyone" can install hardware on the network, the 'security' is already broken so badly I had to use scare quotes.
or
b) go to the company, literally add code on a case-by-case basis to log a particular set of user's actions. This could include real-time alerts, if necessary. Oh, and the FBI is either the one doing the coding, or they pay standard rates for the service's programmers to do the job. This, again, avoids the security issue implicit to a government-mandated backdoor, by moving the "backdoor" from the computer level to the organizational level. It also does privacy better than a), because by being in the application layer instead of the network layer, it can be smart enough to only log the suspected users, not everyone.
This seems totally reasonable. The FBI gets the data they need (face it, there are always going to be times when they're justified in listening in on "private" communications), the internet companies only have to do anything if there's actually enough of a case for a warrant, there's no backdoors for a hacker to exploit, and, if the judges do their job right, everyone's privacy is maintained unless there's enough evidence to justify violating it.
And thus, by being at least mostly reasonable, it is guaranteed to not happen this way.
Most $1 earplugs I could find were at most NRR 33 - they reduce sound by 33 decibels, which in this case would still be 120dB - on the threshold of immediate hearing damage.
It would be smart to also grab some $15 earmuffs (also NRR 30), bringing the noise down to 90dB or so - comparable to highway traffic.
The terrorists must really be loving this. They don't even have to do anything anymore to get the public terrorized - the 'security forces' are taking care of that for them! They don't even have to make half-credible bomb threats anymore - the {Ministry|Department} of Defense will just make threats up for them!
Let's perform a little experiment. Go get a grant for a couple hundred grand. Don't worry, I'll wait.
Next, go out and buy the most modern automobile you can, with as many safety features as you can find. Only restrictions are "must be street-legal" and "must be available to the general public". No military tanks, no experimental Google self-driving cars, nothing like that.
Now get in, find yourself a nice bit of highway. Get up to 60mph/100kmph or so, standard "cruising speed".
Now point yourself straight at the nearest immovable object. A nice, big tree, or perhaps a brick wall.
If your car didn't magically seize control of the vehicle, apply the brakes and take evasive action as necessary, I suppose it must not have had "sufficient safety mechanisms". But, last I checked, "the pilots tried to turn the airplane into a dirtplane" is not something a safety feature can always stop. A good warning system can alert the pilots that they're about to hit a mountain, but even then, the pilots may not have had time to respond, or may not have heeded the warning.
I'm working, in my "free time", on a video game. I put the game data in CSV files, because hey, that means I can edit item values and such in Excel.
I'm seriously considering moving them to a SQLite "database" for performance reasons - my hand-crafter CSV parser is extremely slow, and it would be far less effort to switch to a simple database than to try to optimize it. The main reason I haven't yet? Ease of access to the data. I'll probably do it once I write a full "editor" for the game (I'll need one anyways, because "level design in Excel" is a phrase only uttered on the fifth and lower levels of Hell), but not before.
Oh, and there's the whole "three or four text fields as composite primary key" thing. That doesn't really work well in databases. And mapping those to numeric IDs is going to be tedious.
Slashdot Mirror
Man, is there even anyone on Slashdot stupid enough to fall for this shit?
Yeah, I can see the morons on Facebook and Youtube being technically incompetent and gullible enough to fall for it, but here? Half of us aren't even running an OS this "software" runs on - hell, I'm pretty sure at least one regular user here is running an OS he wrote in his free time.
Talk about "zero return on investment".
Video game graphics could probably benefit from this. Very few people will notice that one pixel is #FA1003 instead of #FC1102, especially when it's replaced 16ms (or, worst-case, 33ms) later with yet another color. It might actually make things "better" - making the rendering seem more analog. Many games are "wasting" power adding film grain or bokeh depth-of-field or lens flares or vignette, to try to simulate the imperfections of analog systems to try to make their graphics less artificial-looking. If you can get a "better" look while using *less* power, all the better.
Actually, I seem to recall hearing about this earlier. For some reason I want to say nVidia specifically has been looking into this.
He linked to a *cable*. And all the reviews are supremely sarcastic or mocking.
He's joking.
This is a little mini-speech I like to give to people, and it's rather appropriate for the this:
In the US, politicians train and study as politicians. They have degrees in political science, or law, or economics, or maybe history or business. Obama was a lawyer. So was Clinton. Bush II had an MBA. Bush I studied economics, as did Reagan. You have to go back all the way to Carter to find a president that had any sort of remotely "practical" training, as a naval officer specializing in nuclear submarines.
In China, politicians train and study as engineers. If shit goes down (as it is wont to do) and the revolution comes, President Hu Jintao could flee to the US, change his name, and live out his life working as an engineer (hydraulic engineering - his first real job was at a hydroelectric plant). Vice President Xi Jinping studied chemical engineering. Premier Wen Jiabao studied geomechanics. Wu Bangguo studied electrical engineering.
Engineering, fundamentally, is "the study of solving problems". It's not, strictly speaking, a science, but an application of science to the real world.
Modern American politics seems to be less about "solving people's problems" and more "making new problems to 'solve' so you can stay in power".
In case you haven't noticed, China is beating us. They're obviously doing something right, and I don't think it's the censorship or the market controls. Their system of government may not be better than ours on paper - slow, central control of everything rarely works for long - but they're better in practice because they have people who actually do the job they're supposed to do.
NOW is our chance. The Chinese seem to be making exactly the mistake we made - their up-and-coming leaders are career politicians, born-and-raised to rule. At the same time, their population boom will be shifting from a worker-heavy populace to a retiree-heavy populace, causing exactly the economic problems we're having now with all the Baby Boomers retiring.
We get one chance to get back on top. We need a government that responds to us, one that works quickly and efficiently for the benefit of everyone.
If anybody knows of any good candidates, speak up. I do not want a lawyer to represent me. I do not want a manager to represent me. I want an engineer, a man (or woman) who solves problems, because we have a lot of problems that need solving.
India: "Hey, has anybody thought we should try controlling the 'net more?"
Korea: "Nah, that's a terrible idea. Maybe a law keeping ISPs from blocking stuff they don't like would be better."
Germany: "Yeah, that sounds good."
Sweden: "Add a clause telling the movie and music companies to stop suing people for more money than some of *us* have, and you'll get my vote!"
Eritrea: "Hear, hear!"
And then the law gets passed and nobody messed with the internet again and we all live happily ever after, the end. ...
Hey, if *they* get to talk about *their* crazy future scenarios, I get to talk about mine.
Does it? I think it's a valid comparison, because it's fundamentally the same sort of situation. Both "services" have both legitimate and illegitimate* uses. Most people would argue that shutting down the freeways would be blatantly wrong, as it harms the vast majority of legitimate users far more than it harms the minority of illegitimate users. So the question then becomes "at what point do you 'shut down' a service that has both uses"? What ratio of illegitimate to legitimate users is necessary? 70%? 50%? 20%?
The obvious "answer" is "when more people use it wrongly than use it rightly", but as with all easy, obvious answers, that is also demonstrably wrong. Take, for another ludicrous example, amphetamines. The illegitimate users vastly outnumber the legitimate, but you'll note that it is not completely banned. Heavily regulated? Yes. Illegal usage punished? Yes. But completely, 100% banned? No, because there are still proper medicinal uses for it. And you can find thousands of other examples, from leaded gasoline to automatic weapons.
So now we've established that a total ban on something with any amount of legitimate use is, at the very least, not an accepted practice. We don't need to rely on abstract philosophical arguments - we can point to concrete examples. So we've essentially "proven" that you should not ban Bittorrent, inasmuch as you can "prove" anything in as loose a field as ethics and law.
So now the question goes from "do we ban Bittorrent?" to "how do we stop illegitimate uses of Bittorrent?", which is the question we really ought to have started with. And that question I'm afraid is too complicated for me to continue delving into.
It's only a bad example if you don't think. Admittedly, getting John Q. Public to actually think may be difficult...
* I'll note also that even "illegitimate" uses of Bittorrent can be legitimate. I've gotten into the habit of torrenting certain games I own, simply because I don't want to be bothered putting the CD every time I want to play. Completely legal, in my case - the 4,999 other people in the swarm may or may not have a similar justification, and I'd probably bet on the "may not".
I have to say, some of their comparisons seem... unfair.
Their main chart compares three things: regular "console connected to display", "current cloud systems", and "GRID cloud rendering".
First off, they cite 66ms latency just at the display level, which is definitely at the higher end of the spectrum. But at least they use the same latency in each.
Their cloud/cloud comparisons are also quite suspicious. Reducing encode by 60%, yeah, I can buy that. Reducing *decode* - which, I remind you, is done client-side - by the same amount is also suspicious in light of their "this does not require an nVidia client, it will work with any h.264 decoder" claim.
Then they claim to have reduced network latency, and significantly (75ms to 30ms). Now, I can vaguely see how they *could* - if they can reduce bandwidth usage significantly, they might eke out a bit less latency, but I highly doubt they can more than *double* their compression efficiency. Unless they're doing something crazy like putting a network interface directly on the GPU (one image they show contradicts that theory), I think this claim is also pretty dubious.
The worst one is the "game pipeline" time. While I can believe that a newer, more powerful graphics card can definitely perform *twice* as well as an older one, I can also state definitively both that "you can put that new card in the home console as well" and "new games will expand to use that power, leaving you back where you started at 100ms render times". The former I can state because they've *already* released Kepler-based cards (to rave review, although my own seems to be backordered); the latter I can state because that's how the industry has worked since at least the late 70s.
Long story short, they seem to be doing some *extremely* unscientific, biased comparisons. Do they probably have something here? Yeah. Is it literally going to be as good as an actual console (or better yet, PC) connected directly to the console, as they claim? No.
Okay, so -gate has been genericized. Happens to almost every word given enough time.
First, it was used for one specific instance: Watergate. Then we had Koreagate, Billygate, Monicagate - any Federal-level political scandal.
Then it went international: Dunagate, Mabelgate, Petrogate - any political scandal
Then it went non-political: Closetgate, Climategate, Cablegate, Crashgate, and those are just the ones that give me an alliteration bonus. Oh, and Bonusgate.
Wikipedia counts one hundred and nineteen distinct -gates beyond the original Watergate (and also excluding fictional -gates). This includes two distinct Memogates, Nannygates, Grannygates, Spygates, Strippergates and three Troopergates. And those are just the ones meeting Wikipedia notability guidelines.
Languages *change*. The news media uses meaningless sound-bite catchphrases. History gets distorted. Get over it.
(I'd like to violate the DMCA, actually, with the business end of a shovel).
Don't worry, that sentiment is mutual.
How so?
Seriously, what features? I've never tried to write anything as low-level as a kernel, but I can't think of much compiler-specific stuff besides a bunch of compiler detection to set up properly-sized typedefs. Maybe differences in assembly syntax?
Also, clang was designed to be a drop-in replacement for gcc in many (not all) ways. Might make it easier for this case.
I've found that code that will compile properly under a variety of compilers tends to be of better quality.
One of my current projects started out on an old 2.x branch of GCC. When I finally got around to updating to a current GCC, I had to fix quite a few bugs before it would actually work - the different compiler was catching problems I hadn't noticed before.
Same when I tried compiling it under Visual Studio, or Clang - the more compilers I made it work under, the less bugs there were in the code.
Now, if a given program actually uses some special feature of GCC, that's fine - if only one compiler will do what you need, that's fine. Or if it's too much work to maintain a "port" - I stopped maintaining the VS project files a while ago, since I no longer used it. But if you have a chance to at least test it against a different compiler, go ahead and give it a shot.
Because it's fun?
Fact: The standards, Unicode in particular, do not specify one or two lines in the "currency symbol". That is left to the font to decide.
I learned this while setting up a currency database. Apparently Brazil (I think it was Brazil) uses the double-barred symbol for *their* currency, and the single-barred for "US Dollars", which are also in relatively common use. Pretty decent idea for distinguishing currencies, IMO - is sure beats US$ vs CA$, or using ISO 4217 codes.
I'm probably a decent subject for a "do the damn kids even know what this stuff is anymore?" experiment - I graduate college in a few weeks. I know, I know, that's a hell of a selection bias right from the start, but let's go with it.
Simple test: go through my smartphone, see if I can ID every icon's source on one particular screen.
First, the phone itself has a handful of icons. There's an arrow pointing backwards for the general "back/undo" button, a series of lines (with the top extended to the left) for "details/menu", a house for "home page", and a magnifying glass for "search". The first two will never really fail, since they were abstract from the start. Houses still have walls, a floor and a roof, although chimneys may be inexplicable in a few generations. Magnifying glasses are still seen at least on TV, where they're heavily associated with "detectives", making "search" a logical connection. Finally, next to the power button are two logos: a vertical line enclosed in a circle, and a padlock. The former has always mystified me, but is widely-used to mean "power device on/off" (usually, but not always, the circle is broken at the top). The latter, however, is still common enough a sight.
Next icons are on the notifications bar. First one of those is "unread voicemails", which I can recognize as an old cassette tape, although I definitely associate that more with its icon use than its physical use, and I will concede that some people my age may never have used one. Then is a square with a triangle in it, indicating that the media player is playing, which is pretty much a purely abstract icon. Then are two different sets of bars for Wifi and mobile status, again purely abstract AFAIK.
Then there's a little image of the phone surrounding by motion lines, indicating "vibrate mode". Since it looks pretty much like the actual device, I can definitely recognize it. Then there's a battery life indicator, and yes, we do know what AAs are. Last notification icon's the "hardest" - an alarm clock, looking like the traditional round analog clocks with the bells on top. I no longer actually use a dedicated alarm clock - I use this phone instead, but the concept of "a distinct device used for keeping time and playing alarms as specified" is not alien to me. Although I never owned an alarm clock shaped like that, I can recognize the round analog "clock", and from there, even without seeing it a million times in pop culture, I could derive "alarm clock" from that.
The main screen has more icons. There's a plain, old-fashioned computer monitor for my terminal-emulator app, a piece of paper and pencil for my text editor app, a spiral-bound calendar for the calendar app, a set of arithmetic symbols on buttons for my calculator app, and so on. The only ones that someone my age may not recognize are:
* the old-style red/blue 3D goggles for the "Google Goggles" app
* the index-card thing for the Contacts app
* the globe icon for the web browser - nobody calls it the "World Wide Web" anymore, so the connection between "world" and "internet" is now a rather loose one.
* and fine, I'll go ahead and list the chess board and knights from my chess app, since I know half the people my age never learned to play chess
Just as a note, on my phone, at least, "Settings" uses a dial, not a set of gears or a wrench. This arguably makes far more sense - a dial is used to change a setting, while gears are what makes something actually work, and a wrench is more used for "assembly" than "configuration".
There's also a few icons that are *new* and still meaningless - can someone explain the Bluetooth logo?
Let's go into one app and see what icons are in *it*. For this experiment, lets try the media player app.
Well, there's a "series of bulleted items" that brings me to the current playlist. Logical. There's two different arrow-based icons for "random" and "repeat", which are also self-explanatory provided you know how abstract arrows work (aside: does anyone know if these abstract ar
This thing is cool and all, don't get me wrong, but the video was a bit of a letdown. Very quiet (when the blog states one goal was "Have lots of moving parts (preferably loud ones)", and not a very good show of how it actually works. It didn't even use 2/3rds of the "digits" - the video stops after reaching 8, so only the first digit of each register was used.
Well, there's a lot of reasons.
First, installing something on the suspect's computer *generally* can't be done stealthily. If the FBI knocked on my door, handed me a warrant, and installed spyware on my computer, I sure as hell won't be doing anything even slightly suspicious on my own computers - a quick hop to the library, or to a friend's house, and then use *theirs*.
They may also not know where the person is. Say they're trying to catch Steve McBadguy, who's "on the run". They know he tends to log in to his Facebook from public computers while he makes his way towards some country with no extradition laws. They could put a tap on every public computer between here and Uzbekistan*, or they could put a tap on one specific Facebook account. I believe I speak for everyone when I say the latter is better for *everyone* - more privacy for us, and less work for the FBI.
* Fact: The United States does not have an extradition treaty with Uzbekistan.
No.
What happens if, say, you get a virus, or your app goes haywire, and The Critical Irreplaceable File gets corrupted or deleted?
RAID protects from hardware failures. It does nothing against software or human errors. And given my history with hardware, software and people, I'd say the first is generally the most reliable.
Look, if it's a data stream, you can record it. I'm not saying everyone should have an API that the FBI can use. I'm not saying we need to record absolutely everything so the FBI can look at it.
What I'm saying is that if the FBI needs to record something and they have enough evidence to get a warrant, they can come in and write their own damn code to log it, we'll put it on the server for as long as the court order says, and then as soon as they're gone we revert the code back to the way it was. Or, the FBI can log every packet themselves, and *they* get the fun task of sifting through billions of TCP packets to find the ones used by Ahmed ibn Badguy.
And if the system *is* anonymous-by-design, well, "that's literally impossible" is generally considered a valid reason to refuse a warrant. I know if the FBI knocked on my door and handed me a warrant for "whatever is 40km beneath the property" and a shovel, I'd call up the judge and tell him that, unfortunately, the laws of science trump even the US Constitution.
Where, exactly, did you get "private business bearing the burden of law enforcement" from what I wrote?
The FBI would either a) install a relatively simple network device themselves, requiring at most a few minutes downtime, b) write some basic logging code themselves, or c) compensate the private-enterprise-programmers for doing (b).
First, the FBI gets a warrant for a particular "wiretap". This should be absolutely mandatory for what I'm about to propose.
Then, off a specific warrant, they go to whichever company the warrant lists, and either:
a) Install a packet-sniffer in front of the web server, logging everything to disk, which is then physically taken by the FBI as evidence - just like a conventional phone wiretap. This avoids the whole "anyone could use the backdoor" - if "anyone" can install hardware on the network, the 'security' is already broken so badly I had to use scare quotes.
or
b) go to the company, literally add code on a case-by-case basis to log a particular set of user's actions. This could include real-time alerts, if necessary. Oh, and the FBI is either the one doing the coding, or they pay standard rates for the service's programmers to do the job. This, again, avoids the security issue implicit to a government-mandated backdoor, by moving the "backdoor" from the computer level to the organizational level. It also does privacy better than a), because by being in the application layer instead of the network layer, it can be smart enough to only log the suspected users, not everyone.
This seems totally reasonable. The FBI gets the data they need (face it, there are always going to be times when they're justified in listening in on "private" communications), the internet companies only have to do anything if there's actually enough of a case for a warrant, there's no backdoors for a hacker to exploit, and, if the judges do their job right, everyone's privacy is maintained unless there's enough evidence to justify violating it.
And thus, by being at least mostly reasonable, it is guaranteed to not happen this way.
Most $1 earplugs I could find were at most NRR 33 - they reduce sound by 33 decibels, which in this case would still be 120dB - on the threshold of immediate hearing damage.
It would be smart to also grab some $15 earmuffs (also NRR 30), bringing the noise down to 90dB or so - comparable to highway traffic.
The terrorists must really be loving this. They don't even have to do anything anymore to get the public terrorized - the 'security forces' are taking care of that for them! They don't even have to make half-credible bomb threats anymore - the {Ministry|Department} of Defense will just make threats up for them!
The terrorists aren't winning. They already won.
Let's perform a little experiment. Go get a grant for a couple hundred grand. Don't worry, I'll wait.
Next, go out and buy the most modern automobile you can, with as many safety features as you can find. Only restrictions are "must be street-legal" and "must be available to the general public". No military tanks, no experimental Google self-driving cars, nothing like that.
Now get in, find yourself a nice bit of highway. Get up to 60mph/100kmph or so, standard "cruising speed".
Now point yourself straight at the nearest immovable object. A nice, big tree, or perhaps a brick wall.
If your car didn't magically seize control of the vehicle, apply the brakes and take evasive action as necessary, I suppose it must not have had "sufficient safety mechanisms". But, last I checked, "the pilots tried to turn the airplane into a dirtplane" is not something a safety feature can always stop. A good warning system can alert the pilots that they're about to hit a mountain, but even then, the pilots may not have had time to respond, or may not have heeded the warning.
Agreed.
I'm working, in my "free time", on a video game. I put the game data in CSV files, because hey, that means I can edit item values and such in Excel.
I'm seriously considering moving them to a SQLite "database" for performance reasons - my hand-crafter CSV parser is extremely slow, and it would be far less effort to switch to a simple database than to try to optimize it. The main reason I haven't yet? Ease of access to the data. I'll probably do it once I write a full "editor" for the game (I'll need one anyways, because "level design in Excel" is a phrase only uttered on the fifth and lower levels of Hell), but not before.
Oh, and there's the whole "three or four text fields as composite primary key" thing. That doesn't really work well in databases. And mapping those to numeric IDs is going to be tedious.
It's not the peak - TFA states the planet's temperature is about 2700C. Which I would call "infernal".
That does, however, explain how the IR emissions are high enough for us to detect here on Earth, light-years away - it's really, really, *really* hot.