Thing is, the CD-ROM 'boom' wasn't the same thing as the dot-com boom, nor is 'Web 2.0'.
The CD-ROM boom was a large number of small companies making crappy software for doing everyday things. Many of them faded out of existence, though the technology market didn't crash.
The dot-com boom and subsequent crash was due to companies forming in order to apply brick-and-mortar business models to the internet-based businesses, thinking that the internet was just a 'virtual storefront', and worked much the same way. Hence your 'online haircut' and 'online pizza delivery' companies.
'Web 2.0' is a model that ensures payment through advertising and subscription to deliver content and allow the users to interact. I'll admit, the most I get to play with Web 2.0 pages is Digg and gmail, but I like to code in the style Web 2.0 implies - quick-response interface rather than static content.
Anyway, Web 2.0 sites are generally more tools for cool things than misplaced business models. There will be some failure, of course - there always is when there's a technology-related paradigm shift in business - but nothing close to the magnitude of the late '90s.
A low resoution DCT comparison would be able to flag frames of a video with some ease, thus fuzzily* filtering out a large amount of negatives. Then, the paid turks sort out what's a positive from what remains.
Mind you, when I say 'DCT', I'm not talking the 8x8 quanta in an MPEG4 stream, I'm talking a single DCT of the whole frame. More computationally intensive, sure, but you get a somewhat scalable data picture that, given a hash that can retain meaningful proximity information, can be compared to other hash tables, and rated on closeness based on how far into the hash tree it goes.
*(ie: what % of frames appear to closely match the hashed frame DCTs of a copyrighted work)
Yeah. All the demand for cheaper stuff means is that Moore's law will apply on the per-dollar level as well as the bleeding edge level - which it is implied to do anyway.
Oh, but this one is good: "True, some distros are more distributed than others. Ubuntu, which is clearly the flavor of the month (who says publicity doesn't matter when it comes to Linux?), is out front of everything else, according to DistroWatch. OpenSUSE, PCLinuxOS, Fedora, and MEPIS round out the top five. (Since this is an enthusiast site, one must assume that Novell and Red Hat are way unrepresented, so one should add those guys into the top tier.)"
I love it, because they fail to notice that OpenSUSE ~= Novell SUSE and RedHat ~= Fedora (~= meaning 'X is the Enterprise version of Y' or 'X is similar enough to Y that they are binary compatible and use the same package management'). I would estimate that Fedora/OpenSUSE have higher numbers because there are a lot more copies of them out there than the 'Business' versions.
I really wanted to say something on the order of 'Oh, wow. I've NEVER heard this argument before...', but there's enough of that going around.
Fact is, if you want to support most linux distros, you only need, max, 6 versions of your program: apt/deb, rpm, yum, emerge, slack/tgz, and linkable blob. In fact you only truthfully need the binary blob version; if your program is worth anything, the community will do the packaging for you.
You know, unless you're an asshole and restrict modified distribution of your 'freely available' program (Adobe, I'm talking to you; you're telling me your license can't allow rearrangement of files so long as the files are unmodified??)
Ok, say you turn the thing on, measure the laser from the short path, and turn it off.
The laser is still in the long path. If you fiddle with it as it exits the long path, the data you measured from the short path should match up (if this weirdness works). If you shut off the machine, your measurements from the short path may or may not still be valid, but since you don't know what happened to the light in the long path, you don't know if it matches up with that of the short path. No paradox there. Just an incompleted experiment.
The issue with gasoline isn't production efficiency, but sustainability.
That 50-100:1 ratio is based on ignoring how the stuff was generated in the first place: ie, biomass formation, and its subsequent breakdown in a geothermal prison.
Yes, I know, natural energy resources get a pass. But the point is how *long* it takes for those natural processes to generate petrol. It's a lot less time than it takes us to drill, process, and ultimately burn it.
Still, with the cellulostic promise of meeting that 100:1 ratio (and the potential for butanol to break right through it), I only have one regret in our energy history.
I'm going to miss Irish Peet when we run out. That stuff smells unlike any other form of lignite.
Corn is actually quite low in sugar per unit plant mass - but we make a LOT of it.
For ethanol production, we actually *malt* corn to get some sugar out of its more starchy bits.
Still, not the point; corn stalks and husks are an otherwise unused resource that, given a good method of converting cellulose to ethanol (or butanol, for that matter), is capable of turning a waste stream into a fuel stream without affecting food supply.
Mandelbrot set, you're a Rorschach test on fire; You're a day-glo pterodactyl You're a heart-shaped box of strings and wire You're one bad-ass fucking fractal!
"The most difficult part of the whole project was installing Windows on Sigbritt's PC," said Jonsson.
Doubtful. (Why even say this? To impress upon people that a high bandwidth connection isn't "hard" to use? Wouldn't the new computer she ostensibly got, since, as the article notes, she's never owned a computer in her life, have come with Windows installed?[1])"
Ah, you're missing something. If she's got a 40Gb connection, chances are that the PC is custom-built to be able to handle that kind of data flow. PCI-E parts, a fibre net card with custom modulation (thanks, Cisco), very likely a metric fuck ton of RAM and a block-striped SATA RAID-0 are all things I would guess would need to be in it to take full advantage of that 40Gb (about 5GByte/s) connection.
My guess is that it was built from parts, and was a pain in the nuts to both install Windows and get all the drivers for the newer hardware running properly.
So, it's the fallacy of micromanagement; they're meddling in the rules to a degree that will harm future business - for quick profits. The quick profits let them keep their jobs and get tasty bonuses, while screwing their eventual successors.
I know what you mean. What's wrong with demanding that doubled bandwidth? As consumers and citizens, bandwidth is of high importance. We have jobs that require varying amounts of bandwidth, and a tiering system leaves out those who can't afford or aren't granted tiered status. I mean, be in the medical industry or have slow uploads? What's a developer to do? FTP takes throughput too, you know.
And debugging webpages isn't a fast process if I have to wait for someone's video surgery to pass through the tubes.
Yes, I went there.
The fact is that you can prioritize bandwidth all you want, but it won't help the impending data traffic jams we're all going to start getting stuck in. We need the infrastructure improved, not preferential treatment.
Good point. If there were an inexpensive nicoteine (slow release) pill I could take, I'd gladly do so rather than smoking. Presently the patch and gum are about as expensive as actual smokes, so I don't see the point.
Mind you, I only want to quit for financial reasons; it's getting expensive to be an addict these days.
Does boolean arithmetic count as math? I would say so.
I dunno. The guy's argument from the article seems kinda flawed.
"A logic circuit is not a sequence of operations."
No, it's a subset of a sequence of operations. It's a component that fits into a deterministic set thereof, and *should* be calculated via boolean or classical arithmetic.
"An operating system is not supposed to terminate, nor does it yield a singular solution."
Then what's "Shut Down" do? And while it doesn't yield a singular solution, it yields a given solution for a given set of inputs.
"An operating system cannot be deterministic because it must relate to uncoordinated inputs from the outside world."
Non-sequitur. Inputs and processes being parallelized does not preclude the individual logical paths from being deterministic, even if the logical paths use each others' states as inputs.
"Any program utilising random input to carry out its process, such...is not an algorithm."
An input does not determine the qualities as an algorithm; even if the random is a preferred or generated input, it's counted as external to the program.
Not to mention that internally generated randoms are algorithmic in nature.
No, seriously. It seems that his entire argument is directed towards changing semantics to take the emphasis off of the mathematical underpinnings of computer science. Rar.
Well, on my laptop, display at 1024x768 (native), SciTE using Bitstream Vera Mono at 9pt gets 139 characters per screen. Drop that to a nice binarily round number, and you get 128, which is generally readable on most screens, given the font is scaled to fit (7pt for 800x600, 11.5 for 1280x1024, 14pt for 1600x1200, etc).
Non Sequitur. I do not willingly distribute my wallet's contents along other channels.
Additional: It's fine if you copy my wallet's contents and give it to your friends. But if you or any of your friends were to use those contents in a way that could be construed as fraud, you can expect to be strung up by your shorts and your curlies.
Related to another comment: Filesharing is the opposite of theft in that you are providing copies of something to others at no cost. In contrast, theft is removing from someone's posession something without paying.
Of course, there *are* costs; bandwidth ain't free, whether it's charged per bit or a monthly fee.
"You just can't expect a program from 1998 to run on your new RedHat system. You can with Windows."
Generally, software from 1998 runs fine on Linux. I mean, if the source hasn't been updated since then, or you can get untouched source from 1998. Just build and run. It's not as if the X Windows framework, Qt, or GTK have changed so far that legacy apps don't work.
Meanwhile, Windows apps from 1998 generally work peachily. So what's the issue here?
"If you can't afford to be upgrading every single time a new major release comes out or run unstable (say you have hundreds of workstations or dozens of servers, for example), then you end up with a bunch of custom compiled apps which actually subverts that packaging system."
I don't see how. Compile once, link many times. That's how this works. As for 'subverting the packaging system', that doesn't seem to make any sense at all. I mean, unless you're compile/building the same apps as you're apt-getting, for example, packagers generally coexist well with compiled. Developers wouldn't have it any other way.
"Now you have to worry about breaking all the custom compiled stuff when you DO finally get round to upgrading the base distribution... which makes the upgrade that much more daunting."
What the hell are you on about? I dunno about Red Hat, but Debian/Ubuntu packages are usually no more than a week behind source.
Generally, any package manager will have something similar to Slackware's checkinstall command to take the last step in building from source. Usually, a given developer needs only produce the source tarball, and the distros, if they find the program useful, will package it up themselves.
For commercial software, yes, they have to do the final link (note: not the compile) separately in every distro; something that will take a couple of hours on a machine with VMWare and the Big Four distros.
Note that VMWare has got this right, from the binblob standpoint; they distribute the necessary object files and have an install script that will do the linking at the host computer, doing all the dependancy checks at link time.
Meanwhile, while there are different paths for common stuff, there's almost always a command to get the path you need for a given thing. (good) Linux developers figure this out while they write their software.
Slashdot Mirror
Thing is, the CD-ROM 'boom' wasn't the same thing as the dot-com boom, nor is 'Web 2.0'.
The CD-ROM boom was a large number of small companies making crappy software for doing everyday things. Many of them faded out of existence, though the technology market didn't crash.
The dot-com boom and subsequent crash was due to companies forming in order to apply brick-and-mortar business models to the internet-based businesses, thinking that the internet was just a 'virtual storefront', and worked much the same way. Hence your 'online haircut' and 'online pizza delivery' companies.
'Web 2.0' is a model that ensures payment through advertising and subscription to deliver content and allow the users to interact. I'll admit, the most I get to play with Web 2.0 pages is Digg and gmail, but I like to code in the style Web 2.0 implies - quick-response interface rather than static content.
Anyway, Web 2.0 sites are generally more tools for cool things than misplaced business models. There will be some failure, of course - there always is when there's a technology-related paradigm shift in business - but nothing close to the magnitude of the late '90s.
Dunno, but I think you're a bit mistaken.
A low resoution DCT comparison would be able to flag frames of a video with some ease, thus fuzzily* filtering out a large amount of negatives. Then, the paid turks sort out what's a positive from what remains.
Mind you, when I say 'DCT', I'm not talking the 8x8 quanta in an MPEG4 stream, I'm talking a single DCT of the whole frame. More computationally intensive, sure, but you get a somewhat scalable data picture that, given a hash that can retain meaningful proximity information, can be compared to other hash tables, and rated on closeness based on how far into the hash tree it goes.
*(ie: what % of frames appear to closely match the hashed frame DCTs of a copyrighted work)
In Soviet New Zealand, Ministers of Parliament ridicule you?
In shorter terms:
These people have a psychosomatic illness caused by their own hypochondria.
Yeah. All the demand for cheaper stuff means is that Moore's law will apply on the per-dollar level as well as the bleeding edge level - which it is implied to do anyway.
Pointless article.
How many sites have the login field on the user generated pages *and* allow users to post javascript?
Few, if any.
Oh, but this one is good:
"True, some distros are more distributed than others. Ubuntu, which is clearly the flavor of the month (who says publicity doesn't matter when it comes to Linux?), is out front of everything else, according to DistroWatch. OpenSUSE, PCLinuxOS, Fedora, and MEPIS round out the top five. (Since this is an enthusiast site, one must assume that Novell and Red Hat are way unrepresented, so one should add those guys into the top tier.)"
I love it, because they fail to notice that OpenSUSE ~= Novell SUSE and RedHat ~= Fedora (~= meaning 'X is the Enterprise version of Y' or 'X is similar enough to Y that they are binary compatible and use the same package management'). I would estimate that Fedora/OpenSUSE have higher numbers because there are a lot more copies of them out there than the 'Business' versions.
I really wanted to say something on the order of 'Oh, wow. I've NEVER heard this argument before...', but there's enough of that going around.
Fact is, if you want to support most linux distros, you only need, max, 6 versions of your program: apt/deb, rpm, yum, emerge, slack/tgz, and linkable blob. In fact you only truthfully need the binary blob version; if your program is worth anything, the community will do the packaging for you.
You know, unless you're an asshole and restrict modified distribution of your 'freely available' program (Adobe, I'm talking to you; you're telling me your license can't allow rearrangement of files so long as the files are unmodified??)
Not exactly.
Ok, say you turn the thing on, measure the laser from the short path, and turn it off.
The laser is still in the long path. If you fiddle with it as it exits the long path, the data you measured from the short path should match up (if this weirdness works). If you shut off the machine, your measurements from the short path may or may not still be valid, but since you don't know what happened to the light in the long path, you don't know if it matches up with that of the short path. No paradox there. Just an incompleted experiment.
Hello, this is Buckaroo Banzai, AKA God.
I'm sending a message back from the end of time when I created the universe.
I hope you're enjoying it.
Well, there are four things the media latches onto when touting the next 'LCD killer':
Faster response time
Light-emission (rather than masking - better efficiencies)
Lower cost
Physical flexibility
Until I see a product with *all four* of these features, I predict LCD will remain the mainstay of thin displays.
The issue with gasoline isn't production efficiency, but sustainability.
That 50-100:1 ratio is based on ignoring how the stuff was generated in the first place: ie, biomass formation, and its subsequent breakdown in a geothermal prison.
Yes, I know, natural energy resources get a pass. But the point is how *long* it takes for those natural processes to generate petrol. It's a lot less time than it takes us to drill, process, and ultimately burn it.
Still, with the cellulostic promise of meeting that 100:1 ratio (and the potential for butanol to break right through it), I only have one regret in our energy history.
I'm going to miss Irish Peet when we run out. That stuff smells unlike any other form of lignite.
Corn is actually quite low in sugar per unit plant mass - but we make a LOT of it.
For ethanol production, we actually *malt* corn to get some sugar out of its more starchy bits.
Still, not the point; corn stalks and husks are an otherwise unused resource that, given a good method of converting cellulose to ethanol (or butanol, for that matter), is capable of turning a waste stream into a fuel stream without affecting food supply.
I actually clicked into this story, pressed 'ctrl+f' and typed 'I for one', just to find your exact comment.
W00t!
Mandelbrot set, you're a Rorschach test on fire;
You're a day-glo pterodactyl
You're a heart-shaped box of strings and wire
You're one bad-ass fucking fractal!
"The most difficult part of the whole project was installing Windows on Sigbritt's PC," said Jonsson.
Doubtful. (Why even say this? To impress upon people that a high bandwidth connection isn't "hard" to use? Wouldn't the new computer she ostensibly got, since, as the article notes, she's never owned a computer in her life, have come with Windows installed?[1])"
Ah, you're missing something. If she's got a 40Gb connection, chances are that the PC is custom-built to be able to handle that kind of data flow. PCI-E parts, a fibre net card with custom modulation (thanks, Cisco), very likely a metric fuck ton of RAM and a block-striped SATA RAID-0 are all things I would guess would need to be in it to take full advantage of that 40Gb (about 5GByte/s) connection.
My guess is that it was built from parts, and was a pain in the nuts to both install Windows and get all the drivers for the newer hardware running properly.
So, it's the fallacy of micromanagement; they're meddling in the rules to a degree that will harm future business - for quick profits. The quick profits let them keep their jobs and get tasty bonuses, while screwing their eventual successors.
I know what you mean. What's wrong with demanding that doubled bandwidth? As consumers and citizens, bandwidth is of high importance. We have jobs that require varying amounts of bandwidth, and a tiering system leaves out those who can't afford or aren't granted tiered status. I mean, be in the medical industry or have slow uploads? What's a developer to do? FTP takes throughput too, you know.
And debugging webpages isn't a fast process if I have to wait for someone's video surgery to pass through the tubes.
Yes, I went there.
The fact is that you can prioritize bandwidth all you want, but it won't help the impending data traffic jams we're all going to start getting stuck in. We need the infrastructure improved, not preferential treatment.
Give me 10M/s, or give me death!
Good point. If there were an inexpensive nicoteine (slow release) pill I could take, I'd gladly do so rather than smoking. Presently the patch and gum are about as expensive as actual smokes, so I don't see the point.
Mind you, I only want to quit for financial reasons; it's getting expensive to be an addict these days.
Does boolean arithmetic count as math? I would say so.
I dunno. The guy's argument from the article seems kinda flawed.
"A logic circuit is not a sequence of operations."
No, it's a subset of a sequence of operations. It's a component that fits into a deterministic set thereof, and *should* be calculated via boolean or classical arithmetic.
"An operating system is not supposed to terminate, nor does it yield a singular solution."
Then what's "Shut Down" do? And while it doesn't yield a singular solution, it yields a given solution for a given set of inputs.
"An operating system cannot be deterministic because it must relate to uncoordinated inputs from the outside world."
Non-sequitur. Inputs and processes being parallelized does not preclude the individual logical paths from being deterministic, even if the logical paths use each others' states as inputs.
"Any program utilising random input to carry out its process, such...is not an algorithm."
An input does not determine the qualities as an algorithm; even if the random is a preferred or generated input, it's counted as external to the program.
Not to mention that internally generated randoms are algorithmic in nature.
No, seriously. It seems that his entire argument is directed towards changing semantics to take the emphasis off of the mathematical underpinnings of computer science. Rar.
Well, on my laptop, display at 1024x768 (native), SciTE using Bitstream Vera Mono at 9pt gets 139 characters per screen. Drop that to a nice binarily round number, and you get 128, which is generally readable on most screens, given the font is scaled to fit (7pt for 800x600, 11.5 for 1280x1024, 14pt for 1600x1200, etc).
Is that why girls keep dumping me. They keep getting smart above the threshold below which is required to tolerate me.
Fun.
This is all very nice, but will my gmail account change hands or names?
Non Sequitur. I do not willingly distribute my wallet's contents along other channels.
Additional: It's fine if you copy my wallet's contents and give it to your friends. But if you or any of your friends were to use those contents in a way that could be construed as fraud, you can expect to be strung up by your shorts and your curlies.
Related to another comment: Filesharing is the opposite of theft in that you are providing copies of something to others at no cost. In contrast, theft is removing from someone's posession something without paying.
Of course, there *are* costs; bandwidth ain't free, whether it's charged per bit or a monthly fee.
"You just can't expect a program from 1998 to run on your new RedHat system. You can with Windows."
Generally, software from 1998 runs fine on Linux. I mean, if the source hasn't been updated since then, or you can get untouched source from 1998. Just build and run. It's not as if the X Windows framework, Qt, or GTK have changed so far that legacy apps don't work.
Meanwhile, Windows apps from 1998 generally work peachily. So what's the issue here?
"If you can't afford to be upgrading every single time a new major release comes out or run unstable (say you have hundreds of workstations or dozens of servers, for example), then you end up with a bunch of custom compiled apps which actually subverts that packaging system."
I don't see how. Compile once, link many times. That's how this works. As for 'subverting the packaging system', that doesn't seem to make any sense at all. I mean, unless you're compile/building the same apps as you're apt-getting, for example, packagers generally coexist well with compiled. Developers wouldn't have it any other way.
"Now you have to worry about breaking all the custom compiled stuff when you DO finally get round to upgrading the base distribution... which makes the upgrade that much more daunting."
What the hell are you on about? I dunno about Red Hat, but Debian/Ubuntu packages are usually no more than a week behind source.
Generally, any package manager will have something similar to Slackware's checkinstall command to take the last step in building from source. Usually, a given developer needs only produce the source tarball, and the distros, if they find the program useful, will package it up themselves.
For commercial software, yes, they have to do the final link (note: not the compile) separately in every distro; something that will take a couple of hours on a machine with VMWare and the Big Four distros.
Note that VMWare has got this right, from the binblob standpoint; they distribute the necessary object files and have an install script that will do the linking at the host computer, doing all the dependancy checks at link time.
Meanwhile, while there are different paths for common stuff, there's almost always a command to get the path you need for a given thing. (good) Linux developers figure this out while they write their software.