I think the one thing that people do *not* think about but should can be summed up in one word: Glass.
It's damn easy to make a zoom lens with a nice zoom rate. It's hard to make a zoom lens that doesn't suffer distortion that gathers a reasonable amount of light, etc.
And it's much harder to quantify a good lens vs. a bad lens in a review in a way that your average clueless folk can compare. So people just go for the numbers and don't bother to think about the glass.
And, similarly, folks buy an expensive SLR and put a really crummy zoom lens on it. When they'd get much much better results with a nice prime lens or two with lots of light gathering ability.
Or, even better, give the spectrum back to the people, who rightfully own it.
Like put some spectrum that's solely devoted to a single family of 802.11-like standards set by the IEEE, with no microwaves, cordless phones, etc. to gunk it up.
Or build allocations around spread-spectrum technologies. Remember, several spread-spectrum networks can exist on the same frequency bands, if done properly, and other networks appear only as noise. If networks are prevented from actively fighting with each other, it's an awfully good way to reward carriers for putting up more access points over a metro area and also to reward more efficent encoding schemes.
The problem is that the spectrum policy made rough sense when you had no choice but to step on other transmissions. And it's lately been seen as a hidden way for the government to make money. But it's not very good for business, nor is it good for us.
First, most of the bigger open-source projects have some sort of funding and support structure. People pay for somebody to do things that *they* want to do and pay for the ability to have somebody come over and fix stuff.
In a sense, if there's enough people who need the same thing, they can cooperate in much the same way as standards are constructed. Remember, open source projects don't have many of the expenses as a pre-packaged concern.
But, also, I think there's a fine line between open-source and you-get-the-source that people sometimes skip over. Meaning, there's not necessarily as much harm in QuarkXPress's source code being on the CD that you purchase as people would like to think.
See, that's the weird bit. Everybody centers around the conspiracy theory that NASA burned the blueprints and forgets that there's many many other reasons for why the Saturn-V is no more. Really, we lost the Saturn-V once all of the mobile launch platforms were cut up to make the shuttle's mobile launch platform.
The thing is, if we need an expendable booster to get 100 tonnes into orbit, we can make one happen, either by strapping engines and a payload canister on the back of a shuttle external tank + SRB or by clustering Atlas V or Delta IV first stages, for much less money than trying to revisit the Saturn V.
In fact, we're closer to having a production-ready heavy lift booster than the Russians are. The Russians would have to restart Energia production or get the Angara booster into production. Energia boosters are very very unlikely to return to production, primarily because the bits are made by different companies in what is now different countries. And the Russians have been developing the Angara boosters very slowly.
Actually, I think the best bit about Bittorrent is that it *does* have legitimate usages that are either street-legal (fansubs, rarities, etc) or fully legal and don't require anonymity and is being used for many of those.
I mean, when everybody slashdotted Scaled Composites' server for a video file, they just put a torrent up.
So SuprNova may end up doing more damage than aid.
Oh, I'm not proud of suffering through crappy teachers. It's not necessary. If I were to rule the world, I'd make sure that teachers were not only able to teach, but to teach you how to learn.
It's just that there's more teachers who are bright but can't teach than teachers who can teach and, at the same time, teach you how to learn properly. So you need to make do with what's available.
Oh yeah, and you are going to have to deal with people who can't communicate every day, including people who have lived in the states every day of their lives.
Unless you know somebody, it's hard to get in to the truly cool jobs. Most companies only recruit at a relatively defined set of universities, generally where the founders and a few of the early employees came from. Which means you have to seek out companies more if you want to avoid being a coding grunt.
Once you are out for a bit, it matters far less.
Oh yeah, and a good CS degree is not about being taught. It's about being tortured into learning because your professor is really bright but can't teach. So he gives you hard tests and you have to teach stuff to yourself in order to pass. At least, that's the shared experience amongst most of the grads from top-10 CS schools that I've talked to.
The problem with half of the possible shuttle missions was that they were presupposed upon the shuttle launching every week.
Repairing a satelite doesn't make sense when the repair mission costs more than a replacement satelite.
So this design makes sense until you get launch costs down. But that's OK, because if you got launch costs down enough, spacecraft construction will be a booming business.
Things that get Really Fsking Hot are black because the only thing that will handle the heat is a carbon-carbon composite.
Things that get Not So Hot are white because it's either that or beige or black when it comes to high temperature ceramics.
There are other alternative coatings like metal, but given that the Russians have already flown a craft with shuttle-like tiles, it's probably the case that they'll stick with those.
Except, of course, that when the Russians coppied the idea of putting tiles on their shuttles, they made them a smidge sturdier.
Paint has undesirable properties, so you want to minimize it's use on the higher-temperature surfaces. If you look at the shuttle, except for small red maintenence markings, they pretty much stuck with that.
Actually, the SDRAM standard allows you to set the memory to self-refresh mode.
I think the biggest problem is that the people who don't care also probably wouldn't want to pay the extra $30 for supercaps and detection logic. And the people who do care want UPS.
So right now I'm using a Motorola v710 on Verizon for my wireless Internet. Using Bluetooth. Dial-up networking is one of the few things that Verizon didn't disable on the v710's Bluetooth stack. I haven't tried it with Linux, but I do know that it's basicly like having a good old Hayes Compatable attached to the computer.
However, that only gets you a maximum 128k theoretical, realistically 56k.
I can give you a link to the AirPrime 5220 card. That can be hacked.
Why the Darwin kernel? Why not the Linux kernel -- it's got the best hardware support. What about the L4 microkernel that has incredibly fast messaging rates?
Why Smalltalk? Why not Ruby or Objective-C or Python? Or maybe ML? All of them are pretty flexible and configurable languages with some nice rapid-development features.
Why the latest graphics technology? What makes a modern graphics system inherrently better in any relatively substantial way, other than sheer bandwidth and 3D rasterization performance.
Mostly, what you need is a clueful UI, which is really independent of any programming properties. You need the OS to not get in the way of blasting data around so that it can play media fast enough. You need the OS to have a low latency so that it can react to input fast enough.
But, most of all, you need useful applications.
All of these things are independent of each other. A crappy kernel, a crappy framework, and bad graphics technology can still have useful applications.
Most metals, if divided finely enough, will burn. Titanium burns *nicely*. Yet you can make the body of a Mach-3 aircraft out of it and there's no problems.
I think the bigger point is that you are more than welcome to sue the ghosts of Infantry's pasts for delivering government-supported beatdowns and the ghosts of many now-dead politicians for their bad decisions. But your ability to compensate for past offenses decreases over time because everybody who was actually hurt as well as anybody related to decisions made is dead. Somebody probably got away with murdering one of my ancestors back in history. If I walk past their descendent now, am I entitled to any sort of compensation? Will I even be able to know that they killed my ancestor long ago?
The thing to remember is that the best-and-brightest of India -- the IIT grads -- do not stay in India, most of them are able to grab green cards and work for even more in the US than they could in India.
The current outsourcing population in India is the second tier. Which is still pretty decent, but there's a limited supply of them and eventually they will price themselves out of the market.
The problem in India is that there's no good third tier. You either have at least bachleurs degree and probably a masters degree, or you are almost illiterate. This isn't any sort of bell-curve intellectual gap, it's mostly that the public pre-college education in India is awful. And there's a lot of waste there -- kids who might become great thinkers but because they are culturally expected to be lower class, they are. I used to think that a good way to disrupt the social order there would be to educate the poor, but it's a much more complicated problem than that.
In the US, at least you still have a good population of folks to send to community college.
Not entirely, and not necessarily for that reason.
Part of the reason for the high cost of living in the big urban cores of the US is not just because there's nothing to do there, but also because there's a lot of demand for everybody in that place.
If you put a large enough population of college educated folks in a hicksville, USA town, you start to have some of the trappings that they expect. Especially if there's already a universty or other similar construction nearby.
No, the part that makes it evil is more that if the one business who really invested in a small town leaves the town for some reason, you have no choice but to move. Which doesn't apply to a big-city based company quite as much.
Closer access to all of the useful services of a city -- like the offices of the companies you work with, service offices, temporary services, etc. The network speed may not be the same for heavy-duty networks. I'm betting that a OC-192 is much easier to get in Silicon Valley than in rural Tenesee.
And also a lot of people who do relatively specilized tasks tend to like being near a cluster of businesses. Even really odd things like being a medical specialist in a town with a single factory in it, where if the factory is shut down, you aren't going to have any business.
It's also a sense of prestige.
Oh yeah, and the afforementioned CEO-gets-a-nice-window-suite advantage.;)
And, even better, the employees have the same sort of legal rights and responsibilities as everybody else.
Which means that abusive office practices aren't as big of a problem for the employees, the employer worries a little less about blatant problems that the police won't touch without a bribe, securities fraud, information leaks, etc.
Slashdot Mirror
I think the one thing that people do *not* think about but should can be summed up in one word: Glass.
It's damn easy to make a zoom lens with a nice zoom rate. It's hard to make a zoom lens that doesn't suffer distortion that gathers a reasonable amount of light, etc.
And it's much harder to quantify a good lens vs. a bad lens in a review in a way that your average clueless folk can compare. So people just go for the numbers and don't bother to think about the glass.
And, similarly, folks buy an expensive SLR and put a really crummy zoom lens on it. When they'd get much much better results with a nice prime lens or two with lots of light gathering ability.
I agree. When I had a 2 day downtime, there was no corresponding reduction in spam.
Or, even better, give the spectrum back to the people, who rightfully own it.
Like put some spectrum that's solely devoted to a single family of 802.11-like standards set by the IEEE, with no microwaves, cordless phones, etc. to gunk it up.
Or build allocations around spread-spectrum technologies. Remember, several spread-spectrum networks can exist on the same frequency bands, if done properly, and other networks appear only as noise. If networks are prevented from actively fighting with each other, it's an awfully good way to reward carriers for putting up more access points over a metro area and also to reward more efficent encoding schemes.
The problem is that the spectrum policy made rough sense when you had no choice but to step on other transmissions. And it's lately been seen as a hidden way for the government to make money. But it's not very good for business, nor is it good for us.
Ahh, but there's a partial counterpoint...
The parts *other* than the set of rules that determine who gets audited are completely harmless.
True.
There's some responses to that, however.
First, most of the bigger open-source projects have some sort of funding and support structure. People pay for somebody to do things that *they* want to do and pay for the ability to have somebody come over and fix stuff.
In a sense, if there's enough people who need the same thing, they can cooperate in much the same way as standards are constructed. Remember, open source projects don't have many of the expenses as a pre-packaged concern.
But, also, I think there's a fine line between open-source and you-get-the-source that people sometimes skip over. Meaning, there's not necessarily as much harm in QuarkXPress's source code being on the CD that you purchase as people would like to think.
Or the contract will just contain a provision prohibiting you from offering a service such as this.
See, that's the weird bit. Everybody centers around the conspiracy theory that NASA burned the blueprints and forgets that there's many many other reasons for why the Saturn-V is no more. Really, we lost the Saturn-V once all of the mobile launch platforms were cut up to make the shuttle's mobile launch platform.
The thing is, if we need an expendable booster to get 100 tonnes into orbit, we can make one happen, either by strapping engines and a payload canister on the back of a shuttle external tank + SRB or by clustering Atlas V or Delta IV first stages, for much less money than trying to revisit the Saturn V.
In fact, we're closer to having a production-ready heavy lift booster than the Russians are. The Russians would have to restart Energia production or get the Angara booster into production. Energia boosters are very very unlikely to return to production, primarily because the bits are made by different companies in what is now different countries. And the Russians have been developing the Angara boosters very slowly.
Actually, I think the best bit about Bittorrent is that it *does* have legitimate usages that are either street-legal (fansubs, rarities, etc) or fully legal and don't require anonymity and is being used for many of those.
I mean, when everybody slashdotted Scaled Composites' server for a video file, they just put a torrent up.
So SuprNova may end up doing more damage than aid.
Oh, I'm not proud of suffering through crappy teachers. It's not necessary. If I were to rule the world, I'd make sure that teachers were not only able to teach, but to teach you how to learn.
It's just that there's more teachers who are bright but can't teach than teachers who can teach and, at the same time, teach you how to learn properly. So you need to make do with what's available.
Oh yeah, and you are going to have to deal with people who can't communicate every day, including people who have lived in the states every day of their lives.
It does matter for your first job.
Unless you know somebody, it's hard to get in to the truly cool jobs. Most companies only recruit at a relatively defined set of universities, generally where the founders and a few of the early employees came from. Which means you have to seek out companies more if you want to avoid being a coding grunt.
Once you are out for a bit, it matters far less.
Oh yeah, and a good CS degree is not about being taught. It's about being tortured into learning because your professor is really bright but can't teach. So he gives you hard tests and you have to teach stuff to yourself in order to pass. At least, that's the shared experience amongst most of the grads from top-10 CS schools that I've talked to.
Dude, the Saturn V Plans aren't lost, it's just incredibly pointless to try and build one at this point. It's an awfully tired myth at this point.
And it doesn't matter because you can launch a mars shot in two or three launches of a shuttle derived booster anyway.
Oops. You are entirely correct.
I found a nice article about the different versions of the thermal protection system.
The problem with half of the possible shuttle missions was that they were presupposed upon the shuttle launching every week.
Repairing a satelite doesn't make sense when the repair mission costs more than a replacement satelite.
So this design makes sense until you get launch costs down. But that's OK, because if you got launch costs down enough, spacecraft construction will be a booming business.
Well, it's actually pretty simple.
Things that get Really Fsking Hot are black because the only thing that will handle the heat is a carbon-carbon composite.
Things that get Not So Hot are white because it's either that or beige or black when it comes to high temperature ceramics.
There are other alternative coatings like metal, but given that the Russians have already flown a craft with shuttle-like tiles, it's probably the case that they'll stick with those.
Except, of course, that when the Russians coppied the idea of putting tiles on their shuttles, they made them a smidge sturdier.
Paint has undesirable properties, so you want to minimize it's use on the higher-temperature surfaces. If you look at the shuttle, except for small red maintenence markings, they pretty much stuck with that.
Actually, the SDRAM standard allows you to set the memory to self-refresh mode.
I think the biggest problem is that the people who don't care also probably wouldn't want to pay the extra $30 for supercaps and detection logic. And the people who do care want UPS.
OK, so I checked out the CDMA Development Group and it's actually 153.6 kbps theoretical max. We're both wrong. ;)
But, yeah, it's user saturation and carrier configuration. You always have your main channel, but you can also have up to 16 suplemental channels.
So right now I'm using a Motorola v710 on Verizon for my wireless Internet. Using Bluetooth. Dial-up networking is one of the few things that Verizon didn't disable on the v710's Bluetooth stack. I haven't tried it with Linux, but I do know that it's basicly like having a good old Hayes Compatable attached to the computer.
However, that only gets you a maximum 128k theoretical, realistically 56k.
I can give you a link to the AirPrime 5220 card. That can be hacked.
One of my honeypot email addresses has received several trojan horse messages from our friends at the spamhausen.
Why does any of that matter?
Why the Darwin kernel? Why not the Linux kernel -- it's got the best hardware support. What about the L4 microkernel that has incredibly fast messaging rates?
Why Smalltalk? Why not Ruby or Objective-C or Python? Or maybe ML? All of them are pretty flexible and configurable languages with some nice rapid-development features.
Why the latest graphics technology? What makes a modern graphics system inherrently better in any relatively substantial way, other than sheer bandwidth and 3D rasterization performance.
Mostly, what you need is a clueful UI, which is really independent of any programming properties. You need the OS to not get in the way of blasting data around so that it can play media fast enough. You need the OS to have a low latency so that it can react to input fast enough.
But, most of all, you need useful applications.
All of these things are independent of each other. A crappy kernel, a crappy framework, and bad graphics technology can still have useful applications.
Actually, they didn't burn so well.
Most metals, if divided finely enough, will burn. Titanium burns *nicely*. Yet you can make the body of a Mach-3 aircraft out of it and there's no problems.
I think the bigger point is that you are more than welcome to sue the ghosts of Infantry's pasts for delivering government-supported beatdowns and the ghosts of many now-dead politicians for their bad decisions. But your ability to compensate for past offenses decreases over time because everybody who was actually hurt as well as anybody related to decisions made is dead. Somebody probably got away with murdering one of my ancestors back in history. If I walk past their descendent now, am I entitled to any sort of compensation? Will I even be able to know that they killed my ancestor long ago?
Actually, that's not necessarily the case.
The thing to remember is that the best-and-brightest of India -- the IIT grads -- do not stay in India, most of them are able to grab green cards and work for even more in the US than they could in India.
The current outsourcing population in India is the second tier. Which is still pretty decent, but there's a limited supply of them and eventually they will price themselves out of the market.
The problem in India is that there's no good third tier. You either have at least bachleurs degree and probably a masters degree, or you are almost illiterate. This isn't any sort of bell-curve intellectual gap, it's mostly that the public pre-college education in India is awful. And there's a lot of waste there -- kids who might become great thinkers but because they are culturally expected to be lower class, they are. I used to think that a good way to disrupt the social order there would be to educate the poor, but it's a much more complicated problem than that.
In the US, at least you still have a good population of folks to send to community college.
Not entirely, and not necessarily for that reason.
Part of the reason for the high cost of living in the big urban cores of the US is not just because there's nothing to do there, but also because there's a lot of demand for everybody in that place.
If you put a large enough population of college educated folks in a hicksville, USA town, you start to have some of the trappings that they expect. Especially if there's already a universty or other similar construction nearby.
No, the part that makes it evil is more that if the one business who really invested in a small town leaves the town for some reason, you have no choice but to move. Which doesn't apply to a big-city based company quite as much.
Closer access to all of the useful services of a city -- like the offices of the companies you work with, service offices, temporary services, etc. The network speed may not be the same for heavy-duty networks. I'm betting that a OC-192 is much easier to get in Silicon Valley than in rural Tenesee.
;)
And also a lot of people who do relatively specilized tasks tend to like being near a cluster of businesses. Even really odd things like being a medical specialist in a town with a single factory in it, where if the factory is shut down, you aren't going to have any business.
It's also a sense of prestige.
Oh yeah, and the afforementioned CEO-gets-a-nice-window-suite advantage.
And, even better, the employees have the same sort of legal rights and responsibilities as everybody else.
Which means that abusive office practices aren't as big of a problem for the employees, the employer worries a little less about blatant problems that the police won't touch without a bribe, securities fraud, information leaks, etc.