Strangely, Linux gave up the "one size fits all" and runs nicely as a desktop O/S on my Fedora Core laptop, and as nicely as a mobile O/S on my Android Moto Droid2 Phone. There is very little software that works on both platforms, they are effectively completely different Operating Systems.
Software engineers like the number 1. Unifying a whole suite of problems into a single framework feels better at a gut level, it just seems right. And even though Microsoft has been trying for almost 20 years to get this unified approach to work and has failed repeatedly, they'll keep trying because they are software engineers of the modest type - the type arrogant enough to think they have all the right answers but not quite smart enough to figure out how nor that it's a bad idea.
Somewhere deep in Silicon Valley, a programmer is looking at a comment something like this:
/******* FIXME: WTF Hack here. CRLs require authentication of being revoked, but we never bothered to check the callback of the revoke. Maybe if we bothered to have a revoke infrastructure? For now, we'll just not bother fixing this until 10.1 or 10.2. ******/ return true;
A patent application must be filed before a technology is made publicly available, or it is no longer patentable. That is already the case even under current patent law, and will not change.
Sorry, not true. You have 1 year from public disclosure to patent the idea.
My software company was in line to provide signature validation services for the State of California. Although we didn't land the contract, finding out what it took to become a legally recognized CA for California was part of the process. California (and by extension, most governments) requires a SAS70 audit. Performed once, and then re-performed annually. The audit itself cost about $25,000, we estimated the actual cost of compliance at $250,000.
That's an approximation of what it costs to become a legally recognized CA.
The hardware/software combination for generating the certificates is $50 for a used computer on EBay and a download of a Linux ISO. Most of the cost isn't in the technology, but the operational processes in making sure the certificates are managed properly.
We just replaced some shema-heavy databases to 4x OCZ Vertex3 120 GB drives about a month ago. The queries are monsters, 12-table joins with combined inner/outer are par for the course, CentOS/PostgreSQL.
We saw similar performance numbers. We didn't get too specific, just ran some rough testing, and in the worst case, we saw at least a 90% drop in time-to-execute critical queries. System load dropped from 1-5 under light/medium loads to 0.5 under heavy loads.
We were a bit worried about the SSDs holding up under sustained load so we spent 4 months pounding on a couple HARD on dev servers - never raised a sweat. The SSDs so thoroughly stomped the SCSI drives in our developer shop that the server with the SSDs instantly became the one ALL our developers connected to simultaneously - and it held up gracefully.
It's damned impressive to have a drive actually max out a 3 Gb SATA interface in random, small-block reads.
The idea of spectrum allocation is, in my opinion, itself an antiquated idea. Instead, we should be using spread-spectrum technology that actually makes use of the *entire* radio spectrum rather than inefficiently chopped up segments.
Really, take a look at the radio charts and you'll find that even in "heavy use" areas like the cities, the actual use of frequencies is actually rather small, with a few frequencies (such as cellular) suffering serious collisions while others are virtually empty.
While I can see allocating a small number of frequencies for emergency services and perhaps aviation, I'd personally be all good with deregulating all the rest and using error-tolerant frequency hopping spread-spectrum algorithms to distribute the traffic load across the entire spectrum. The result would be something like a 100,000 fold increase in available bandwidth.
Trusting a typical performance benchmark is like trusting now many pounds you can bench as a measure of overall health. Yes, benchmarks do provide some information about the relative performance of products, but it's by no means a complete picture!
When you are evaluating performance of a system, it's important to consider all aspects of system performance, and not get too hung up on any single factor.
Recently, we did a server upgrade to using SSDs in our database servers. The performance difference was dramatic, though in benchmark testing, we saw very little performance difference between 1.5 Gbps SATA interfaces and 6 Gbps SATA III interfaces for hosting production databases. We saw something like a 90% drop in system load! This indicates that the biggest benefit of SSDs was in reduced seek time, not transfer rate, and that for our needs, transfer speeds weren't a bottle neck in any event, but that the sharply reduced seek times provided a huge benefit.
Know what you need, and you be much more likely to get it!
The concept is known as trust - presuming that previous experience will continue. If Apple started churning out rubbish products that didn't fulfil peoples needs and expectations, then they might last one generation, but they would soon lose their reputation.
Macs were initially populare, but faded until they became an expensive grey box that performed badly *cough* Performa */cough*, which continued right up to the point where Apple almost went belly up.
Enter: The Steve! He slashed the whole lineup and introduced the iMac - a cute, stylish, decent performing computer with the oh-so-sexy, Unix-based OSX!
Followed by the iPod, iTunes, the iPhone, and now the iPad, all quality products with obvious, early attention given in the pretotype stage so that the initial experience was as good as the ongoing one.
... sayeth somebody who lives under the umbrella of one of the most successful, free-society-fostering governments on Earth.
Yes, our government has its share of flaws, and we should work to correct them. But it's idiocy to claim that one of the most successful systems ever devised should be unilaterally dismissed.
I remember writing a steganographic tool that sent hidden messages via SPAM. I had a massive source of SPAM and use a combination of hash tables and a psuedo-random number generator in order to pass any type of binary data as SPAM. The trick was to have all possible combinations of spam words with offsets for all the hexidecimal characters. Numbers are scarce, but spam is such bad spelling that you could "cheat" a little and get it all set up. (My favorite trick was to embed a meaningless tag that had the hexidecimal value I needed in it)
I'm sure the NSA could have cracked my little toy / experiment fairly quickly, but they would only have cracked the fact that I was using steganography. The binary stream encoded therein could still easily be encrypted with AES or Blowfish or any other cryptomechanism.
The stream expansion was pretty intense, some 20:1 or so. But it was honest-to-god SPAM and it was fun to cleanly pass compressed, encrypted binary data via penis pill offers.
When growing food crops one is generally left with a large amount of left over cellulose - which is why it is generally cheap. We use it to feed cattle, as fertilizer - but really we just want to get rid of it. Being able to use this cellulose for fuel production would be a huge help and would not have an impact on food production.
I live in the Central Valley, California. Every fall, the sky is smokey and smoggy because of all the burning of rice fields. Hundreds of square miles of rice fields produce a massive amount of food for the world, and the end result is rice stalks - too woody to be eaten by livestock, it takes more than a year to decompose, yet it's not woody enough to make a good building material, though numerous attempts have been made to concoct some sort of usable fiberboard out of it.
So we burn it. All that horrid, dreadful cellulose!
Why, particularly? I would guess that which one would be "better" would be a calculation that combines ease of access, cost, aesthetics, and ROI. Often, operations done at large scale can be done more efficiently than in a distributed fashion. Other times, the cost of distribution can offset this interent efficiency.
We don't yet know which one is "better" - the market is still merging.
One area that I'd personally love to see more solar panels is over parking lots. Nothing quite beats the misery of walking out of a nice, 75 degree mall into the blistering, 100-degree heat in the summer time, only to sit down in your 160 degree car, cursing and swearing at all that damned free energy the sun packed into your car.
But cover that parking lot with a lattice of solar panels so I'm getting into a merely hot 95 degree car while all that energy is used to power the A/C at the mall I just got out of, that would be swell.
This same thing can happen with SSH, FTP, and any other service that uses password authentication.
There. 'Nuff said. Passwords are terrible for system level security and should not be used. The basic idea of passwords requires that, to use it, you also give everything needed to use it again. Techniques like two-channel authentication, public key encryption, etc. solve this problem.
Google is never perfect, in general it is only significantly better then the alternative, and it forces the alternative to work for the better.
For *decades*, this would have been an apt description of Microsoft. In the early days, *nix was a disjointed, proprietary mess dominated by sneering, long-haired elitists, computers were out of reach of average people, and costs for everything was high.
Microsoft has gradually ossified, and now stands in a strange position: it seems to want to be a startup, but its actual revenue betrays the fact that most of their income comes from just two products: Windows and Office. Worse, it's been that way for a long time, and none of their many and numerous attempts to regain a lead position anywhere else have panned out.
You could consider Xbox an exception, except that it's really just a PC with Windows' DirectX pre-installed. (Yes, we're back to Windows)
It would be pretty kick ass to be able to take the train all of the way from Boston to London, by way of Canada, the US, Russia, Latvia, Lithuania, Poland, Germany, Belgium, and France...
kick ass, I guess.... if you don't mind the 2 or 3 weeks that you'd spend sitting in a rail car. Don't get me wrong! I support the use of efficient rail anywhere it makes sense, but passenger trains is not that place. Airplanes will do the same trip in a single night flight, at rather sharply reduced cost. Only if you are looking at CO2 footprint do trains start to measure up to trains for travel, and then, it would still probably be close / cheaper to travel from Boston to London by plane than by train. Boat wins here, I'm guessing, for a "green route".
Where this has a chance is as an alternate freight shipping route vis-a-vis supertankers, and they are not exactly environmentally fantabulous.... while trains are practically angels in comparison.
It also has the fastest growth of the recent smartphone app stores for published applications.
90% growth! From 1 MS funded app to 10 in just 6 months could achieve this statistic!
Android is here, has the buzz, and has the goods to back it. MS no longer has to be "good enough", they have to be compellingly better. They had everything they needed to make it happen, including a decade (yes, a DECADE) to figure it out with WinMo 1 through WinMo 6.x. After all those generations, they still had only a cheesy interface that vaguely resembled Windows 3.1.
They had all the opportunity in the world, and they managed to blow it trying to bring the "PC experience" to mobile devices, despite the market spending 10 years letting them know that they didn't want the "PC experience".
MS will probably have to buy Google to put this genie back in the bottle....
As stated elsewhere, there are various ways around this limitation, including the use of reversible computing which works by "borrowing" entropy resulting in an extremely low entropy computation mechanism.
Ttight security is *expensive*. Special electronic keys (and the associated administration) costs thousands of dollars. Few organizations actually have sufficiently valuable data to justify this expense. Smaller organizations rarely do.
So far, so good. The tough part is that as an organization grows, it reaches a point where it will start to make sense to incorporate these additional expenses. But what triggers this decision? There's no automatic formula involved, and a growing organization has to carefully track expenses so that it maintains sufficient capital to fund additional growth. And that's where potential breaches such as the "Verizon guy" in the blue suit and a box on a hand truck make it all the way into the server room.
As my company has grown, I've made it a point to secure our assets in (relatively) secure areas. Important servers locked in steel server closets are the norm now, with only two key trusted staff having keys. This level of security is relatively tight, but doesn't scale well. This works well for now because our company is still not that large.
PS: Using high density rackmounts you can pack a surprising amount of horsepower in a single, locked, 42U rack!
Basically having GPS in my pocket at all times has made my business trips far less stressful!
For me, it's ALL kinds of trips! Even towns I kinda know are more pleasant when I can just get to the nearest bank branch, grocery store, fast food joint, hotel, or whatever. It lets me get the inane stuff out of the way without worry, so I have more time and physical/mental energy to enjoy the experience itself.
Also, a drive is much more pleasant when I can queue up an impromptu TED talk or stream something I actually like when I'm stuck in country-music-and-spanish-only radio station territory on a long drive.
And then, when I get a nice shot next to the Bay or under the airplane wing, I can instantly share it with my family and friends. This, too is often a nice way to enjoy the experience even further.
My Android smart phone/GPS isn't just nice, it's a game-changer.
Vacuum tubes produce a unique sound that digital doesn't replicate
I'd like to see *any study whatsoever* that demonstrates that this effect is anything but a placebo. Show me any case *at all* where somebody can consistently discern the difference between the two, all else being the same.
Guess what: choices are EXPENSIVE. All by themselves. Choices themselves have a high price. Fewer options reduces the cost of choosing, and makes you more likely to buy. Witness Apple, which has the "consumer" model and the "pro" model... and that's it! Did you want the Macbook or the Macbook pro?
Sorry, it's just true. Too many choices mean people leave.
There was a resource starvation issue found in apache a while back. Basically, if you give a long post size in the headers of a form post, and then start sending 1 byte every per seconds, the process would remain open. You could then launch 100,000 more such processes and crash the server.
At the time I wrote a quick PHP script as a proof of concept and was able to bring my dev server to its knees in less than 10 minutes, including coding time.
You don't have to be particularly good, just good enough.
There's what's legal, and what's likely to be prosecuted. If you install optware bind on your dd-wrt Buffalo router it's not like men in black will bust down your door. Laws like this are directed at commercial providers and they provide compliance for 98% of the populace.
Commercial providers have their revenue stream to protect, so they comply with laws like this with minimal oversight. What you do in your living room is pretty much up to you. (with a few exceptions)
Slashdot Mirror
Strangely, Linux gave up the "one size fits all" and runs nicely as a desktop O/S on my Fedora Core laptop, and as nicely as a mobile O/S on my Android Moto Droid2 Phone. There is very little software that works on both platforms, they are effectively completely different Operating Systems.
Software engineers like the number 1. Unifying a whole suite of problems into a single framework feels better at a gut level, it just seems right. And even though Microsoft has been trying for almost 20 years to get this unified approach to work and has failed repeatedly, they'll keep trying because they are software engineers of the modest type - the type arrogant enough to think they have all the right answers but not quite smart enough to figure out how nor that it's a bad idea.
Somewhere deep in Silicon Valley, a programmer is looking at a comment something like this:
/*******
FIXME: WTF Hack here. CRLs require authentication of being revoked, but we never bothered to check the callback of the revoke. Maybe if we bothered to have a revoke infrastructure? For now, we'll just not bother fixing this until 10.1 or 10.2.
******/
return true;
A patent application must be filed before a technology is made publicly available, or it is no longer patentable. That is already the case even under current patent law, and will not change.
Sorry, not true. You have 1 year from public disclosure to patent the idea.
Silly engineer-type! You expect sentences to make sense?
Here's what normal people hear: blah blah blah job creation blah blah reform blah blah landmark legislation blah blah better future!
Sadly, people with this level of comprehension have a vote that counts just as much as yours. Welcome to America.
My software company was in line to provide signature validation services for the State of California. Although we didn't land the contract, finding out what it took to become a legally recognized CA for California was part of the process. California (and by extension, most governments) requires a SAS70 audit. Performed once, and then re-performed annually. The audit itself cost about $25,000, we estimated the actual cost of compliance at $250,000.
That's an approximation of what it costs to become a legally recognized CA.
The hardware/software combination for generating the certificates is $50 for a used computer on EBay and a download of a Linux ISO. Most of the cost isn't in the technology, but the operational processes in making sure the certificates are managed properly.
We just replaced some shema-heavy databases to 4x OCZ Vertex3 120 GB drives about a month ago. The queries are monsters, 12-table joins with combined inner/outer are par for the course, CentOS/PostgreSQL.
We saw similar performance numbers. We didn't get too specific, just ran some rough testing, and in the worst case, we saw at least a 90% drop in time-to-execute critical queries. System load dropped from 1-5 under light/medium loads to 0.5 under heavy loads.
We were a bit worried about the SSDs holding up under sustained load so we spent 4 months pounding on a couple HARD on dev servers - never raised a sweat. The SSDs so thoroughly stomped the SCSI drives in our developer shop that the server with the SSDs instantly became the one ALL our developers connected to simultaneously - and it held up gracefully.
It's damned impressive to have a drive actually max out a 3 Gb SATA interface in random, small-block reads.
The idea of spectrum allocation is, in my opinion, itself an antiquated idea. Instead, we should be using spread-spectrum technology that actually makes use of the *entire* radio spectrum rather than inefficiently chopped up segments.
Really, take a look at the radio charts and you'll find that even in "heavy use" areas like the cities, the actual use of frequencies is actually rather small, with a few frequencies (such as cellular) suffering serious collisions while others are virtually empty.
While I can see allocating a small number of frequencies for emergency services and perhaps aviation, I'd personally be all good with deregulating all the rest and using error-tolerant frequency hopping spread-spectrum algorithms to distribute the traffic load across the entire spectrum. The result would be something like a 100,000 fold increase in available bandwidth.
Trusting a typical performance benchmark is like trusting now many pounds you can bench as a measure of overall health. Yes, benchmarks do provide some information about the relative performance of products, but it's by no means a complete picture!
When you are evaluating performance of a system, it's important to consider all aspects of system performance, and not get too hung up on any single factor.
Recently, we did a server upgrade to using SSDs in our database servers. The performance difference was dramatic, though in benchmark testing, we saw very little performance difference between 1.5 Gbps SATA interfaces and 6 Gbps SATA III interfaces for hosting production databases. We saw something like a 90% drop in system load! This indicates that the biggest benefit of SSDs was in reduced seek time, not transfer rate, and that for our needs, transfer speeds weren't a bottle neck in any event, but that the sharply reduced seek times provided a huge benefit.
Know what you need, and you be much more likely to get it!
The concept is known as trust - presuming that previous experience will continue. If Apple started churning out rubbish products that didn't fulfil peoples needs and expectations, then they might last one generation, but they would soon lose their reputation.
Macs were initially populare, but faded until they became an expensive grey box that performed badly *cough* Performa */cough*, which continued right up to the point where Apple almost went belly up.
Enter: The Steve! He slashed the whole lineup and introduced the iMac - a cute, stylish, decent performing computer with the oh-so-sexy, Unix-based OSX!
Followed by the iPod, iTunes, the iPhone, and now the iPad, all quality products with obvious, early attention given in the pretotype stage so that the initial experience was as good as the ongoing one.
... sayeth somebody who lives under the umbrella of one of the most successful, free-society-fostering governments on Earth.
Yes, our government has its share of flaws, and we should work to correct them. But it's idiocy to claim that one of the most successful systems ever devised should be unilaterally dismissed.
I remember writing a steganographic tool that sent hidden messages via SPAM. I had a massive source of SPAM and use a combination of hash tables and a psuedo-random number generator in order to pass any type of binary data as SPAM. The trick was to have all possible combinations of spam words with offsets for all the hexidecimal characters. Numbers are scarce, but spam is such bad spelling that you could "cheat" a little and get it all set up. (My favorite trick was to embed a meaningless tag that had the hexidecimal value I needed in it)
I'm sure the NSA could have cracked my little toy / experiment fairly quickly, but they would only have cracked the fact that I was using steganography. The binary stream encoded therein could still easily be encrypted with AES or Blowfish or any other cryptomechanism.
The stream expansion was pretty intense, some 20:1 or so. But it was honest-to-god SPAM and it was fun to cleanly pass compressed, encrypted binary data via penis pill offers.
When growing food crops one is generally left with a large amount of left over cellulose - which is why it is generally cheap. We use it to feed cattle, as fertilizer - but really we just want to get rid of it. Being able to use this cellulose for fuel production would be a huge help and would not have an impact on food production.
I live in the Central Valley, California. Every fall, the sky is smokey and smoggy because of all the burning of rice fields. Hundreds of square miles of rice fields produce a massive amount of food for the world, and the end result is rice stalks - too woody to be eaten by livestock, it takes more than a year to decompose, yet it's not woody enough to make a good building material, though numerous attempts have been made to concoct some sort of usable fiberboard out of it.
So we burn it. All that horrid, dreadful cellulose!
Why, particularly? I would guess that which one would be "better" would be a calculation that combines ease of access, cost, aesthetics, and ROI. Often, operations done at large scale can be done more efficiently than in a distributed fashion. Other times, the cost of distribution can offset this interent efficiency.
We don't yet know which one is "better" - the market is still merging.
One area that I'd personally love to see more solar panels is over parking lots. Nothing quite beats the misery of walking out of a nice, 75 degree mall into the blistering, 100-degree heat in the summer time, only to sit down in your 160 degree car, cursing and swearing at all that damned free energy the sun packed into your car.
But cover that parking lot with a lattice of solar panels so I'm getting into a merely hot 95 degree car while all that energy is used to power the A/C at the mall I just got out of, that would be swell.
This same thing can happen with SSH, FTP, and any other service that uses password authentication.
There. 'Nuff said. Passwords are terrible for system level security and should not be used. The basic idea of passwords requires that, to use it, you also give everything needed to use it again. Techniques like two-channel authentication, public key encryption, etc. solve this problem.
Google is never perfect, in general it is only significantly better then the alternative, and it forces the alternative to work for the better.
For *decades*, this would have been an apt description of Microsoft. In the early days, *nix was a disjointed, proprietary mess dominated by sneering, long-haired elitists, computers were out of reach of average people, and costs for everything was high.
Microsoft has gradually ossified, and now stands in a strange position: it seems to want to be a startup, but its actual revenue betrays the fact that most of their income comes from just two products: Windows and Office. Worse, it's been that way for a long time, and none of their many and numerous attempts to regain a lead position anywhere else have panned out.
You could consider Xbox an exception, except that it's really just a PC with Windows' DirectX pre-installed. (Yes, we're back to Windows)
It would be pretty kick ass to be able to take the train all of the way from Boston to London, by way of Canada, the US, Russia, Latvia, Lithuania, Poland, Germany, Belgium, and France...
kick ass, I guess.... if you don't mind the 2 or 3 weeks that you'd spend sitting in a rail car. Don't get me wrong! I support the use of efficient rail anywhere it makes sense, but passenger trains is not that place. Airplanes will do the same trip in a single night flight, at rather sharply reduced cost. Only if you are looking at CO2 footprint do trains start to measure up to trains for travel, and then, it would still probably be close / cheaper to travel from Boston to London by plane than by train. Boat wins here, I'm guessing, for a "green route".
Where this has a chance is as an alternate freight shipping route vis-a-vis supertankers, and they are not exactly environmentally fantabulous.... while trains are practically angels in comparison.
It also has the fastest growth of the recent smartphone app stores for published applications.
90% growth! From 1 MS funded app to 10 in just 6 months could achieve this statistic!
Android is here, has the buzz, and has the goods to back it. MS no longer has to be "good enough", they have to be compellingly better. They had everything they needed to make it happen, including a decade (yes, a DECADE) to figure it out with WinMo 1 through WinMo 6.x. After all those generations, they still had only a cheesy interface that vaguely resembled Windows 3.1.
They had all the opportunity in the world, and they managed to blow it trying to bring the "PC experience" to mobile devices, despite the market spending 10 years letting them know that they didn't want the "PC experience".
MS will probably have to buy Google to put this genie back in the bottle....
As stated elsewhere, there are various ways around this limitation, including the use of reversible computing which works by "borrowing" entropy resulting in an extremely low entropy computation mechanism.
Ttight security is *expensive*. Special electronic keys (and the associated administration) costs thousands of dollars. Few organizations actually have sufficiently valuable data to justify this expense. Smaller organizations rarely do.
So far, so good. The tough part is that as an organization grows, it reaches a point where it will start to make sense to incorporate these additional expenses. But what triggers this decision? There's no automatic formula involved, and a growing organization has to carefully track expenses so that it maintains sufficient capital to fund additional growth. And that's where potential breaches such as the "Verizon guy" in the blue suit and a box on a hand truck make it all the way into the server room.
As my company has grown, I've made it a point to secure our assets in (relatively) secure areas. Important servers locked in steel server closets are the norm now, with only two key trusted staff having keys. This level of security is relatively tight, but doesn't scale well. This works well for now because our company is still not that large.
PS: Using high density rackmounts you can pack a surprising amount of horsepower in a single, locked, 42U rack!
-MMC/SD/eMMC doesn't come close to the throughput of SATA
Not that your other points don't have merit, but the OCZ Vertex 3 bumps up against the throughput limits of 6 Gbps SATA. Next time you might not want to make a point that's countered in the summary, unless the summary is just wrong.
Start with ... SSDs that are fast approaching the 6Gbps ceiling imposed by the current Serial ATA specification...
Basically having GPS in my pocket at all times has made my business trips far less stressful!
For me, it's ALL kinds of trips! Even towns I kinda know are more pleasant when I can just get to the nearest bank branch, grocery store, fast food joint, hotel, or whatever. It lets me get the inane stuff out of the way without worry, so I have more time and physical/mental energy to enjoy the experience itself.
Also, a drive is much more pleasant when I can queue up an impromptu TED talk or stream something I actually like when I'm stuck in country-music-and-spanish-only radio station territory on a long drive.
And then, when I get a nice shot next to the Bay or under the airplane wing, I can instantly share it with my family and friends. This, too is often a nice way to enjoy the experience even further.
My Android smart phone/GPS isn't just nice, it's a game-changer.
Vacuum tubes produce a unique sound that digital doesn't replicate
I'd like to see *any study whatsoever* that demonstrates that this effect is anything but a placebo. Show me any case *at all* where somebody can consistently discern the difference between the two, all else being the same.
Too many choices, is like too much money.
that's the fiction we'd all like to believe. Unfortunately, it's just not true. The cost of too many choices is no choice at all. Yes, it's a TED talk, and it's one of the best I've ever listened to.
You think your choices are up to you?
Guess what: choices are EXPENSIVE. All by themselves. Choices themselves have a high price. Fewer options reduces the cost of choosing, and makes you more likely to buy. Witness Apple, which has the "consumer" model and the "pro" model... and that's it! Did you want the Macbook or the Macbook pro?
Sorry, it's just true. Too many choices mean people leave.
There was a resource starvation issue found in apache a while back. Basically, if you give a long post size in the headers of a form post, and then start sending 1 byte every per seconds, the process would remain open. You could then launch 100,000 more such processes and crash the server.
At the time I wrote a quick PHP script as a proof of concept and was able to bring my dev server to its knees in less than 10 minutes, including coding time.
You don't have to be particularly good, just good enough.
There's what's legal, and what's likely to be prosecuted. If you install optware bind on your dd-wrt Buffalo router it's not like men in black will bust down your door. Laws like this are directed at commercial providers and they provide compliance for 98% of the populace.
Commercial providers have their revenue stream to protect, so they comply with laws like this with minimal oversight. What you do in your living room is pretty much up to you. (with a few exceptions)