Put it out there, and people will find a use for it
You can do serious computing with 64kB (second generation mainframes and 16-bit minicomputers) and 300 baud modems. It can be done. When that was all there was, that is what was done. But with current technology and pricing, it's not what one would chose to do. What we now take for granted, and how we use it was not discernable from what we had then. Similarly we cannot expect to be able to see what use will be made, but we can make some educated guesses.
Multiple remote backups. Cheap proactive disaster planning. There are two aspects to security. Keeping others out of private stuff. This gets almost all of the current attention, but much more critical is not losing your own stuff. If you make your house burglar-proof and lock yourself out of it permantently, you have severe problems. (If you're rich enough you can solve the problem by having several houses;) As an oversimplification, the system of backups needs about ten times the storage at about ten percent of the budget. A lot of open source achieves this with mirrors, including unofficial mirrors left on peoples' hard drives. One problem with backup systems is that they tend to faithfully replicate errors in the main system. There are a few ways a system can work correctly. The ways a system can fail or work incorrectly is essentially unbounded. Several cheap fast independent concurrent backups seems to be the only practicable means of protection. It also helps immensely if you know what you do have and what you do not have. Isuspect many expensive and elaborate backup systems are extremely brittle.
I'd guess that a lot of the attempts to use the bandwidth will be more or less successful attempts at home entertainment systems, however any such has a very real problem. The main draw for the crowds at Mardi Gras are the crowds at Mardi Gras. There's a big difference between live and canned.
essentially ruining any chances of you giving away the software for free and surviving off support calls.
You give away sofware for free, without support, without even phone number. You charge for supported software, and both you and the customer prefer that there doesn't need to be any support calls. However, when and if the customer does run into trouble is not the right time to start to establish a relationship. The customer really needs as much of that stuff as is practicable done before the customer runs into trouble. You really want to make the money from support that is available and not needed (yet). The customer also gains from this kind of relationship.
You (as a customer) have something that is supported and you spend half a day mostly on hold? Everybody loses in that scenario.
Really? That's pretty pessimistic and short sighted...after all, that's what the Windows Update panel, your virus software tray control and your spybot summary are...control panels that tell you whether or not your machine has any problems they can detect.
Hmmmm, NT4 Workstation, IE5, unpatched for 2 or 3 years, maybe more. No Windows Update panel. No Virus software. No spybot summary. I run as root (domain admin) and my machines stay up and logged in. Poor security and poor attitude, yes. But I have my limits. A false sense of security has never been a good idea, and something like a "Security Control Panel" is too much like painting a target for any passing malware.
Most customers aren't interested in the details of software development, they just want a product that meets their need and someone they can complain to, if they're in trouble.
To oversimplify, they aren't buying the software, they're buying a phone number and some sort of access/support that's vaguely related to the price. Regardless of whatever happens to get written into some contract, The vendor wants to keep good customers and the customer wants to keep good dependable vendors. Non-paying customers? will get some support if and to the extent they have "interesting" problems. Ultimately though it boils down to Read The Fine Manual. No manual? Then Write The Fine Manual. Well-paying customers will get support even when the problems are mundane and no longer interesting. That's really what they are paying for.
"What if somebody could tell if their machine was secure just by opening a control panel?" That's a very good idea -- [Emphasis added.]
Only if the intention to extended insecurity as long as possible. Mind game. Imagine this on OpenBSD. You know it's a gag, just like BSOD screensavers.
Does anyone else think it's funny that slowing the spread of viruses is incompatible with Windows?
Situation Normal All Fxxxed Up.
Just one of the ways you know Microsoft has never been serious about security. The best security is when your own people are aware of what's going on and are in a position to put a stop to stuff going on that shouldn't be going on. The problem's not getting a virus, it's passing it on to a bunch of unsuspecting friends and neighbors.
tendency in our current society to call pretty much everyone who isn't a sheep a "terrorist."
The thing is that we as a society need people who are "different". Even if they were mostly wrong, we need them. Otherwise we get too inbred, like people who keep marrying first cousins.
since MS has a near total monopoly on PC software, any suggestion for the use of software that is non-MS becomes "Anti-microsoft" for the simple reason that this is what the user must be switching FROM.
Any suggestion for the use of non-MS software is "Anti-Microsoft". Any realistic evaluation of various software is "Anti-Microsoft". (That excluded funded by Microsoft studies) Any attempts by users to regain their computers from Microsoft must be "Anti-Microsoft".
The joys of a monopoly. Whatever the problems with current computers, blame it on Microsoft. Legitimately. Microsoft is the only one really in a position to do anything about it.
The whole point is that they don't see that they are getting burned, no matter what happened. Most of them don't care even when they do get burned, a situation even worst than you described.
Always Blame Microsoft. First, blame Microsoft. Then find out why. Surprisingly effective. It's not your fault. It's not my fault. It's Microsoft's fault.
In a word, no. The horse is dead when the worms stop coming.
The current main page of Slashdot is a good place to find out what's happening now. And not just for regulars. I don't mind dupes. I read/. quite regularly but many times I've never seen the first round.
The culture determines the language, not vice versa
Methinks it a feedback loop, but that point is very well made. Seems like the French language is the way it is because that's the way that French culture wants it to be rather than French culture is an accident of the French language. I'm no expert on French, but do know enough that given a choice between being right and being French, they will make the obvious choice.
The culture determines the language. The culture expresses itself in that language. The culture uses the language to distinguish between acceptable and unacceptable behavior. Analysis is made extremely difficult when the carrier of information is what is not said. When a distinction needs to be made the culture will find some means of expressing that distinction. When a distinction is unimportant the culture will ignore that apparent distinction. This can make foreign languages difficult because sounds that ought to be different are the same and sounds that ought to be the same are different.
Also, there is not just one culture in operation, there are several somewhat overlapping cultures, each making its own set of distinctions. Medicine seems to like the prefixes hyper and hypo where the distinction is in the unaccented vowel, thoroughly confusing the average layman.
Stop trying to claim that forks are bad and stop trying to claim they don't happen under the GPL.
I don't claim that forks are bad. I don't claim that forks do not happen under the GPL. Quite the contrary. The various distributions will push a bit in some direction or another in an attempt to distinguish themselves. Where these are actually successful they will show in a few generations in the other distributions. It's what happens after the forking that matters. While the GPL certainly does not require merging stuff, there is certainly that tendency in the long term.
If someone starts a magazine company that provided useful Business information with NO bias I'll subscribe.
The Economist. Someone else said it best to the effect that if you have in depth knowledge of a subject, The Econominst doesn't do bad things to your digestion.
NO bias. Unlikely, the writers seem to want to be biased on the side on insight, which tends to require some bias. But fat chance figuring out which "side" they're on.
BSD has a few forks. Linux has many forks. Main-line linux, as on kernel.org Red Hat Linux, with a modified (forked) kernel. Various patch sets that haven't (yet) made it into the main line. By applying different combinations of patch sets, you can have more different possible kernels, this is before you start configuring, than there are places to put those kernels.
The critical difference is that Red Hat 2.6 is a fork from the main-line 2.6 not a 2.6 fork of Red Hat 2.4. Forks that aren't worth keeping up with just wither and die. Linux gives the outward appearance of not forking, because there is about as much merging as there is forking.
The GPL doesn't prevent forking Round 1. True. Round 2. Pretty much true. Round 3. Significant amount of merging of Round 1 forks. Round 4. More merging.
With a constant forking rate, GPL should eventually reach a smallish number of forks. The GPL does prevent someone from holding turf just because of a few improvements. If they're actually good, in general, "everybody" will have them.
The "Unix Wars" scenario is a result of everybody finding some way, any way, to make their product distinctive and different. There was a time when I think everybody managed to do something different in the arrangement of the keys on a keyboard. But strange is not usually better. You get the same thing the first time a user gets a word processor with lots of different fonts.
Every success is dogged by advice to abandon the wellspring of that success.
"Don't do so good. You make the rest of us look bad."
Artists tend to become a lot more famous after they're dead. A dead artist has significantly more problems contradicting the art critics than a live artist.
This shows that they were mentally rotating the character in a languageless manner, then deciding if it were mirrored or not.
The time was related to the difficulty of perception. Once the distinction is perceived, and the process of this perception can be assumed to be more fundamental than verbalized language, the distinction itself is cast into language, ie mirrored or not mirrored.
Try physical manipulations with grunts and noises, and I don't mean exasperation. There's a language there.
1000 machines are more reliable then 1 big machine.
Depends. If any machine(s) are up, the system works. This is Google's advantage. If any machine hiccups, the system fails. This is supercomputer turf.
A cluster is a fast efficient way to solve a few problems. A very small portion of the problems would be my guess. To the extent that the next computation depends on something non-local, which in turn depends in part on the results of your last computation, supercomputers have an advantage somewhat like the difference between cache and disk access speeds.
I've always remembered it as "discriminatory pricing". I assume by "sell below costs" you mean below average costs, not incremental costs. Of course with software and mirrors, incremental costs can be negligible. The interesting thing about discriminatory pricing is that you can have a very healthy and profitable market where there is no single price for everybody where the market is viable.
Slashdot Mirror
Put it out there, and people will find a use for it
You can do serious computing with 64kB (second generation mainframes and 16-bit minicomputers) and 300 baud modems. It can be done. When that was all there was, that is what was done. But with current technology and pricing, it's not what one would chose to do. What we now take for granted, and how we use it was not discernable from what we had then. Similarly we cannot expect to be able to see what use will be made, but we can make some educated guesses.
Multiple remote backups. Cheap proactive disaster planning.
There are two aspects to security. Keeping others out of private stuff. This gets almost all of the current attention, but much more critical is not losing your own stuff. If you make your house burglar-proof and lock yourself out of it permantently, you have severe problems. (If you're rich enough you can solve the problem by having several houses;)
As an oversimplification, the system of backups needs about ten times the storage at about ten percent of the budget. A lot of open source achieves this with mirrors, including unofficial mirrors left on peoples' hard drives.
One problem with backup systems is that they tend to faithfully replicate errors in the main system. There are a few ways a system can work correctly. The ways a system can fail or work incorrectly is essentially unbounded. Several cheap fast independent concurrent backups seems to be the only practicable means of protection. It also helps immensely if you know what you do have and what you do not have. Isuspect many expensive and elaborate backup systems are extremely brittle.
I'd guess that a lot of the attempts to use the bandwidth will be more or less successful attempts at home entertainment systems, however any such has a very real problem. The main draw for the crowds at Mardi Gras are the crowds at Mardi Gras. There's a big difference between live and canned.
essentially ruining any chances of you giving away the software for free and surviving off support calls.
You give away sofware for free, without support, without even phone number.
You charge for supported software, and both you and the customer prefer that there doesn't need to be any support calls. However, when and if the customer does run into trouble is not the right time to start to establish a relationship. The customer really needs as much of that stuff as is practicable done before the customer runs into trouble. You really want to make the money from support that is available and not needed (yet). The customer also gains from this kind of relationship.
You (as a customer) have something that is supported and you spend half a day mostly on hold? Everybody loses in that scenario.
Really? That's pretty pessimistic and short sighted...after all, that's what the Windows Update panel, your virus software tray control and your spybot summary are...control panels that tell you whether or not your machine has any problems they can detect.
Hmmmm, NT4 Workstation, IE5, unpatched for 2 or 3 years, maybe more. No Windows Update panel. No Virus software. No spybot summary. I run as root (domain admin) and my machines stay up and logged in. Poor security and poor attitude, yes. But I have my limits. A false sense of security has never been a good idea, and something like a "Security Control Panel" is too much like painting a target for any passing malware.
Most customers aren't interested in the details of software development, they just want a product that meets their need and someone they can complain to, if they're in trouble.
To oversimplify, they aren't buying the software, they're buying a phone number and some sort of access/support that's vaguely related to the price. Regardless of whatever happens to get written into some contract, The vendor wants to keep good customers and the customer wants to keep good dependable vendors. Non-paying customers? will get some support if and to the extent they have "interesting" problems. Ultimately though it boils down to Read The Fine Manual. No manual? Then Write The Fine Manual. Well-paying customers will get support even when the problems are mundane and no longer interesting. That's really what they are paying for.
Linus accepting a job at SCO: 1,000,000:1
Intermediate odds. Linus buys SCO for $1.00.
"What if somebody could tell if their machine was secure just by opening a control panel?" That's a very good idea -- [Emphasis added.]
Only if the intention to extended insecurity as long as possible.
Mind game. Imagine this on OpenBSD. You know it's a gag, just like BSOD screensavers.
8 times, easily.
E-sized drafting paper.
Not that thin either.
Trial by fire.
Supported? I've seen claims of "supported" where the "support" was documentation to disable the feature. (Closed Source can be that bad;)
Not Supported? Might work but you don't even get sympathy if it doesn't.
One thing might help at least some of us. How the expletive-deleted do you tell what model of what it is that you've got in your hands?
In any event, the best of luck to you. This could get very interesting.
Does anyone else think it's funny that slowing the spread of viruses is incompatible with Windows?
Situation Normal All Fxxxed Up.
Just one of the ways you know Microsoft has never been serious about security.
The best security is when your own people are aware of what's going on and are in a position to put a stop to stuff going on that shouldn't be going on. The problem's not getting a virus, it's passing it on to a bunch of unsuspecting friends and neighbors.
tendency in our current society to call pretty much everyone who isn't a sheep a "terrorist."
The thing is that we as a society need people who are "different". Even if they were mostly wrong, we need them. Otherwise we get too inbred, like people who keep marrying first cousins.
since MS has a near total monopoly on PC software, any suggestion for the use of software that is non-MS becomes "Anti-microsoft" for the simple reason that this is what the user must be switching FROM.
Any suggestion for the use of non-MS software is "Anti-Microsoft".
Any realistic evaluation of various software is "Anti-Microsoft". (That excluded funded by Microsoft studies)
Any attempts by users to regain their computers from Microsoft must be "Anti-Microsoft".
The joys of a monopoly. Whatever the problems with current computers, blame it on Microsoft. Legitimately. Microsoft is the only one really in a position to do anything about it.
The whole point is that they don't see that they are getting burned, no matter what happened. Most of them don't care even when they do get burned, a situation even worst than you described.
Always Blame Microsoft.
First, blame Microsoft. Then find out why.
Surprisingly effective.
It's not your fault.
It's not my fault.
It's Microsoft's fault.
Prepare the ground for "There is a better way."
Havent we talked about this enough allready?
/. quite regularly but many times I've never seen the first round.
In a word, no.
The horse is dead when the worms stop coming.
The current main page of Slashdot is a good place to find out what's happening now. And not just for regulars. I don't mind dupes. I read
The culture determines the language, not vice versa
Methinks it a feedback loop, but that point is very well made.
Seems like the French language is the way it is because that's the way that French culture wants it to be rather than French culture is an accident of the French language. I'm no expert on French, but do know enough that given a choice between being right and being French, they will make the obvious choice.
The culture determines the language. The culture expresses itself in that language. The culture uses the language to distinguish between acceptable and unacceptable behavior. Analysis is made extremely difficult when the carrier of information is what is not said. When a distinction needs to be made the culture will find some means of expressing that distinction. When a distinction is unimportant the culture will ignore that apparent distinction. This can make foreign languages difficult because sounds that ought to be different are the same and sounds that ought to be the same are different.
Also, there is not just one culture in operation, there are several somewhat overlapping cultures, each making its own set of distinctions. Medicine seems to like the prefixes hyper and hypo where the distinction is in the unaccented vowel, thoroughly confusing the average layman.
Stop trying to claim that forks are bad and stop trying to claim they don't happen under the GPL.
I don't claim that forks are bad.
I don't claim that forks do not happen under the GPL.
Quite the contrary.
The various distributions will push a bit in some direction or another in an attempt to distinguish themselves. Where these are actually successful they will show in a few generations in the other distributions.
It's what happens after the forking that matters. While the GPL certainly does not require merging stuff, there is certainly that tendency in the long term.
If someone starts a magazine company that provided useful Business information with NO bias I'll subscribe.
The Economist.
Someone else said it best to the effect that if you have in depth knowledge of a subject, The Econominst doesn't do bad things to your digestion.
NO bias. Unlikely, the writers seem to want to be biased on the side on insight, which tends to require some bias. But fat chance figuring out which "side" they're on.
BSD has a few forks.
Linux has many forks.
Main-line linux, as on kernel.org
Red Hat Linux, with a modified (forked) kernel.
Various patch sets that haven't (yet) made it into the main line.
By applying different combinations of patch sets, you can have more different possible kernels, this is before you start configuring, than there are places to put those kernels.
The critical difference is that Red Hat 2.6 is a fork from the main-line 2.6 not a 2.6 fork of Red Hat 2.4. Forks that aren't worth keeping up with just wither and die. Linux gives the outward appearance of not forking, because there is about as much merging as there is forking.
You beat me to it, but I'll add this.
The GPL doesn't prevent forking
Round 1. True.
Round 2. Pretty much true.
Round 3. Significant amount of merging of Round 1 forks.
Round 4. More merging.
With a constant forking rate, GPL should eventually reach a smallish number of forks. The GPL does prevent someone from holding turf just because of a few improvements. If they're actually good, in general, "everybody" will have them.
The "Unix Wars" scenario is a result of everybody finding some way, any way, to make their product distinctive and different. There was a time when I think everybody managed to do something different in the arrangement of the keys on a keyboard. But strange is not usually better. You get the same thing the first time a user gets a word processor with lots of different fonts.
Every success is dogged by advice to abandon the wellspring of that success.
"Don't do so good. You make the rest of us look bad."
Artists tend to become a lot more famous after they're dead. A dead artist has significantly more problems contradicting the art critics than a live artist.
Monday Morning Quarterbacking.
SOP for a new CEO.
If it was centralized then decentralize it.
If it was decentralized then centralize it.
This shows that they were mentally rotating the character in a languageless manner, then deciding if it were mirrored or not.
The time was related to the difficulty of perception.
Once the distinction is perceived, and the process of this perception can be assumed to be more fundamental than verbalized language, the distinction itself is cast into language, ie mirrored or not mirrored.
Try physical manipulations with grunts and noises, and I don't mean exasperation. There's a language there.
The corporations and the big networks have sucked all the joy of the Olympics.
And the excitement and the interest.
Too controlled.
Boring. Boring.
if you can get a big enough cluster that will get the work done faster than the supercomputer and still be cheaper
The problem is that for the kinds of problems where you need a supercomputer, the bigger the cluster, the slower the cluster.
Clusters work well where, given a system of n equations in n unknowns, you can solve the n equations "one equation at a time".
1000 machines are more reliable then 1 big machine.
Depends.
If any machine(s) are up, the system works. This is Google's advantage.
If any machine hiccups, the system fails. This is supercomputer turf.
A cluster is a fast efficient way to solve a few problems. A very small portion of the problems would be my guess. To the extent that the next computation depends on something non-local, which in turn depends in part on the results of your last computation, supercomputers have an advantage somewhat like the difference between cache and disk access speeds.
I've always remembered it as "discriminatory pricing".
I assume by "sell below costs" you mean below average costs, not incremental costs. Of course with software and mirrors, incremental costs can be negligible.
The interesting thing about discriminatory pricing is that you can have a very healthy and profitable market where there is no single price for everybody where the market is viable.