GNU/Linux is at a strange place in it's adoption cycle, and this is a real concern: By the time you are savvy enough with computers to think outside of the marketing and go with Linux as an easy, usable operating system that does everything a beginning user does - you're no longer a beginning user and probably have some application - productivity, gaming, whatever - for which there is no Linux equivalent. As a long-time Linux user, mostly for its traditional core roles as a server and personal development workstation, I must say this is one of the most reasonable arguments against Linux I've seen.
I do quite a bit of work in the GIMP, Skencil, Inkscape, Scribus, and K3d, and I'll admit they're not Photoshop, Quark, and Bryce. They're not nearly as far behind the curve as the video tools, though. Cinelerra is certainly no FCP, or even Final Cut Express. I tend to use Avid's free trial edition on Windows, which is actually enough for my simple needs. I'd love to see Avid or Apple put out their even entry-level stuff on Linux.
What hardware support does Apple have that RedHat and Canonical don't? They specifically have a narrow range of supported hardware approved by Apple. You'd have a much harder time finding hardware to stick into a Mandriva, Ubuntu, Debian, RedHat, Fedora, or Mint system that doesn't have drivers than you would sticking random hardware into a Mac under OS X.
Most of the software you need is available from the repository under mainstream Linux distros. Most other things have an installer, an RPM or DEB package, or a tar.gz archive with installation instructions a third grader could follow available elsewhere. Try getting half of that software running on OS X. Sure, there's a better shrink wrapped software market for OS X, but check out project announcements on Freshmeat sometime. There's plenty of software for most people that runs on Linux.
If you have a higher-end later-model G4 then it'll run Leopard. My 500Mhz AGP graphics Mac is stuck on Tiger. 500Mhz on a G4 is a serious amount of computing power for its day, but its day seems to be past.
The Mac Mini is great for an office desktop maybe, but for a home PC or any serious graphics work I'm going to need better graphics and quite possibly more storage.
"Appearance of the hardware" is about as important as "shiny", even if "shiny" isn't the part of the appearance of the hardware people are concerned about. If it's ugly and it works, stick it under your desk instead of showing it off on top.
Well done! You'd be surprised how many Americans can't seem to figure that out. They are particularly, oddly enough, often the Democrats who are bemoaning Bush in every way but can't see that spending money we don't have is causing lots of the problems.
BTW, our Congress is 49% Democract, 49% Republican, 2% independents who tend to vote Democrat. Our House of Representatives is 56% Democrat. Guess what body of the government passed the budgets according to the President's whims? Yep, the Democrat-controlled Congress, which has an even lower satisfaction rating (16%-22% by many sources) than Bush.
Many of us in the US have been pushing for the minor parties to gain ground on the Democrat/Republican duopoly, since they're pretty much working together towards mostly common goals while appearing to be at one another's throats.
"Can't" is a strong word, but a powered parachute tends to descend very slowly when power is cut. They're usually only built for one or two people, though, and not much else.
As a veteran of the ISP business from the days of 14.4kbps up through ISDN, DSL, and tower-to-home wireless, let me assure you the little players didn't have the funds to do DPI on everyone.
For the places I worked, if we suspected one or two particular customers were into kiddie porn or trying to break our network, we could divert that user's packets through a path with a fairly high level of scrutiny. That was a really rare thing, and we were usually too busy keeping the servers and networking equipment running to bother even when we had suspicions.
We never, at any company I worked for, turned over information on which customer was on an IP address at a given time without a subpoena unless we had proof internally of either kiddie porn or attacking our networks. Any other crime, and the cops had to get their court order. One company I worked for even put up challenges to some subpoenas.
I'm pretty sure the upstream providers for those small to medium ISPs could carry on vastly more streams of packet inspection, but not enough to snoop on everyone. They might've done random sampling or investigated on a suspicion-led basis as we did. The tech to do DPI on everyone I think is just now becoming economically feasible. Although it probably could have been done with the tech of five years ago, I doubt anyone would have paid the cost.
I can't wait til I can get a tap of gasoline into my sink for the type of savings I get on tap water over bottled.;-)
Seriously, though. Bottled water at $1.50 a liter? My wife and I spend $25-30 a quarter year on tap water for drinking, washing dishes, washing clothes, showers, baths, steaming the carpets, water gun fights, flushing toilets, and washing household surfaces. I've paid far more in other cities, but still nowhere near $1.50 a liter. I guess it helps that I'm using a city-owned bulk distribution network and I live along one of the world's biggest rivers.
If we had the same price discrepancy on gasoline, it'd be cheaper to generate our own electricity from that instead of using the grid. Hell, we'd probably try to figure out safe ways to cook with it, heat with it, and dilute it to use as a cleaning solvent.
They're not really that different. A monopoly, oligopoly, or cartel just artificially limits supply.
Here are different inputs to supply and demand: subsidies artificially increase supply (and therefore often artificially increase demand), monopolies artificially limit supply (and sometimes artificially limit demand due to the high price, in favor of margins over total revenue), embargoes artificially limit supply (which often leads to black and gray markets), price caps artificially increase demand (by keeping the deterrent to purchase out of the way), excises artificially lower demand, tariffs restrict both supply and demand in the importing country and raise supply in the exporting country, production limits artificially reduce supply and sometimes psychologically induce increased demand (think collectibles like Beanie Babies)
Nothing really breaks the laws of supply and demand. There are many ways, though, to finagle them.
In several parts of the US, there is the option to get a carrier that doesn't charge for incomding texts. U.S. Cellular is one of those, at least on most plans.
This depends a great deal on your carrier. U.S. Cellular is mine, and things are a bit different.
I don't pay for incoming calls or incoming texts. I get 900 minutes to anywhere in the US, nights between 2100 and 0900 free, weekends from 2100 Fridays to 0900 Monday free, free calls to everyone on the same network, free calls to my voicemail from anywhere in the US, roaming's included across the US and parts of Canada, and I get unlimited Net data use on the phone. I pay about $71 a month including taxes and fees.
There are downsides, too, though. My unlimited data only works through a WAP proxy, and they want $4.99 more a month to give me access to a real web browser not limited to their WAP proxy with limited content. My data plan is limited to use by the phone for the phone, and isn't for use as a modem by a laptop. It's possible for me to set up my own WAP proxy for the phone, though, and I've played with one as a toy but I found I really only check what's on their limited plan from my phone anyway. Others might want to read whole full-size websites from their phone, but not me. I'd really like if they let me use the phone for my PC's data needs for that $10 part of my plan, but of course that's not going to happen.
The towers were there long before anybody thought of SMS. Not only that, but the software to handle SMS messages is a slight tweak to that already used for voice. And the cellular voice software was basically tweaked from terrestrial hard-line phone service only in that it tracks your phone for tower-to-tower roaming.
The hardware, real estate (although lower than rights-of-way for every hardline) and electricity costs are real. Nearly all the technological advances came from existing landlines, packet radio, and the Internet more so than being invented from scratch.
Hell, CDMA/Edge is very close to being a roaming wireless Ethernet and TDMA/GSM is close to being a roaming wireless Token Ring or FDDI. The AT command set of GSM phones comes from modems. The voice data transmitted across a cell network is encoded with audio codecs just like for landlines, audio files on computers, or VoIP (although often using different codecs specifically designed for cellular telephony).
Every cellphone is basically a cellular radio WAN bridge, a NIC, a CPU/DSP combination encoding and decoding packet data, and a user interface for getting sound, numbers, and text into and out of the system. With sufficient standards, cheap enough spectrum, and enough interoperability testing we could easily have Open Source/Open Hardware phones running commodity hardware and software on commodity networks like chat software on PCs across the Internet.
Spectrum's not cheap, though, and interoperability isn't always easy. Google's play to get the open access rules in the recent US spectrum auction is a big step towards a wireless Internet extravaganza, but only a step.
Laziness of people keeping them from grabbing a real computer and instead opting for holding on to the little phone has a lot to do with it. IMs are zero marginal cost once you have Internet service. SMS usually cost quite a bit.
Just use a laptop (or phone handset) with WiFi and a cable/DSL/fiber/tower-to-customer wireless Internet connection at the other end of that WiFi and you can save a bunch of money.
Sure, in some situations/locations, you get SMS without getting a decent Net connection, but I see people sitting on couches or in coffee shops using SMS.
You could print letters in gold. It's been done for centuries. Getting the gold into a fine powder that's suspended in a solvent that sprays through your inkjet and doesn't void your manufacturer's warranty, now that's a trick.
Well, "going round to my mate's house" for the friend I call the most means driving for 45 minutes or more. I think I'll stick with calling when all I want to say is "what's up?".
The (monetary) efficiency of the final payload is exactly what interests the author. That so much more data is used to send a paltry 160 characters is part of thew reason SMS fairs so poorly from a cost standpoint.
Whether it's avoidable considering Hubble always sends the data to the same few places vs. the needed routing for millions of cell phones to send to millions of other cell phones is a fair question, though.
One way the cost per byte of final payload can be lowered is to allow more bytes of payload per message so the percentage of transmitted data that is overhead shrinks. However, in the case of SMS, we get into the cost factor for the hardware, battery, radio, and software in the handsets to try to change that ratio.
You make a point, but it turns out his project isn't another on-line messaging system. I'm not sure how much competition there is with his project, but it's surely less than IM apps. It's an n-body simulation system.
First, I disagree fully that the suggestion of Mechanical Turk is "Offtopic" for this discussion. Some moderator seems to have gotten trigger happy.
A little bit of super cheap by-the-feature labor might be useful. It might help to recruit someone who's interested in the problem. It might help attract more interest as a project when the updates are flowing from more than one contributor. Replace MT contributors with volunteers as they become available if you like. Take donations from users who don't have sufficient programing skills to pay for the MT work.
I've never used the particular service myself, so I'm not sure what quality you'd get. It's entirely possible, in theory, for that to be a viable option though.
I agree, but it's probably better at this point if he can get the domain-aware users ready to program than to get the CS students up to speed on his subject.
Which open source and closed source companies makes a big difference here. Which companies you consider one or the other does, too.
IBM, Sun, Adobe, Corel, Novell, and Apple do a bit of both. Red Hat is considered an open source company, and their books look pretty good. Google uses Open Source software and gives part of their code improvements back even though they're using the code internally. SCO is a closed source company, as are Borland, Fog Creek Software, IMSI, Opera Software, and thousands more. Lernout & Hauspie was a closed-source company, and they are now gone (leaving the Dragon Software founders in the lurch, actually, after buying that company out with stock that ended up unsellable). For every Microsoft, HP, IBM, Sage, or Adobe there are thousands or tens of thousands of little software companies.
One market where OSS really helps is in the really small freelance programmer and software consulting segment. Many of these companies don't have a packaged product to sell, but sell custom programming to multiple clients. They can use OSS as a starting point and customize it rather than building a huge library of software from scratch. This gets smaller software companies lots of business earlier in the operation of their companies, because they can undercut the closed-source competition. The money might be spread out more instead of being in one person's pockets, but most programmers aren't the chairman of Microsoft, and are happy to make a decent living doing what they do.
Since most of the software in the world isn't sold in shrink-wrapped packages, you might want to consider the importance of custom development.
I do quite a bit of work in the GIMP, Skencil, Inkscape, Scribus, and K3d, and I'll admit they're not Photoshop, Quark, and Bryce. They're not nearly as far behind the curve as the video tools, though. Cinelerra is certainly no FCP, or even Final Cut Express. I tend to use Avid's free trial edition on Windows, which is actually enough for my simple needs. I'd love to see Avid or Apple put out their even entry-level stuff on Linux.
It looks like Avid is looking for a Principal Software Engineer" for C/C++ on Linux, so hopefully that'll signal the starter's pistol for the great Linux NLVE race.
Hardware/software compatibilities?
What hardware support does Apple have that RedHat and Canonical don't? They specifically have a narrow range of supported hardware approved by Apple. You'd have a much harder time finding hardware to stick into a Mandriva, Ubuntu, Debian, RedHat, Fedora, or Mint system that doesn't have drivers than you would sticking random hardware into a Mac under OS X.
Most of the software you need is available from the repository under mainstream Linux distros. Most other things have an installer, an RPM or DEB package, or a tar.gz archive with installation instructions a third grader could follow available elsewhere. Try getting half of that software running on OS X. Sure, there's a better shrink wrapped software market for OS X, but check out project announcements on Freshmeat sometime. There's plenty of software for most people that runs on Linux.
I can build a PC in less time than you spend insulting people on Slashdot in day from what I can tell.
Show me a mainstream peripheral that uses firewire800. Plenty of PC machines have firewire400, but most peripherals stop at 400 or use USB2 instead.
If you have a higher-end later-model G4 then it'll run Leopard. My 500Mhz AGP graphics Mac is stuck on Tiger. 500Mhz on a G4 is a serious amount of computing power for its day, but its day seems to be past.
No. There's the part about "decent specs".
The Mac Mini is great for an office desktop maybe, but for a home PC or any serious graphics work I'm going to need better graphics and quite possibly more storage.
"Appearance of the hardware" is about as important as "shiny", even if "shiny" isn't the part of the appearance of the hardware people are concerned about. If it's ugly and it works, stick it under your desk instead of showing it off on top.
Well done! You'd be surprised how many Americans can't seem to figure that out. They are particularly, oddly enough, often the Democrats who are bemoaning Bush in every way but can't see that spending money we don't have is causing lots of the problems.
BTW, our Congress is 49% Democract, 49% Republican, 2% independents who tend to vote Democrat. Our House of Representatives is 56% Democrat. Guess what body of the government passed the budgets according to the President's whims? Yep, the Democrat-controlled Congress, which has an even lower satisfaction rating (16%-22% by many sources) than Bush.
Many of us in the US have been pushing for the minor parties to gain ground on the Democrat/Republican duopoly, since they're pretty much working together towards mostly common goals while appearing to be at one another's throats.
"Can't" is a strong word, but a powered parachute tends to descend very slowly when power is cut. They're usually only built for one or two people, though, and not much else.
You also tend to fly a little more efficient route than a car would travel, resulting in fewer miles traveled.
As a veteran of the ISP business from the days of 14.4kbps up through ISDN, DSL, and tower-to-home wireless, let me assure you the little players didn't have the funds to do DPI on everyone.
For the places I worked, if we suspected one or two particular customers were into kiddie porn or trying to break our network, we could divert that user's packets through a path with a fairly high level of scrutiny. That was a really rare thing, and we were usually too busy keeping the servers and networking equipment running to bother even when we had suspicions.
We never, at any company I worked for, turned over information on which customer was on an IP address at a given time without a subpoena unless we had proof internally of either kiddie porn or attacking our networks. Any other crime, and the cops had to get their court order. One company I worked for even put up challenges to some subpoenas.
I'm pretty sure the upstream providers for those small to medium ISPs could carry on vastly more streams of packet inspection, but not enough to snoop on everyone. They might've done random sampling or investigated on a suspicion-led basis as we did. The tech to do DPI on everyone I think is just now becoming economically feasible. Although it probably could have been done with the tech of five years ago, I doubt anyone would have paid the cost.
I can't wait til I can get a tap of gasoline into my sink for the type of savings I get on tap water over bottled. ;-)
Seriously, though. Bottled water at $1.50 a liter? My wife and I spend $25-30 a quarter year on tap water for drinking, washing dishes, washing clothes, showers, baths, steaming the carpets, water gun fights, flushing toilets, and washing household surfaces. I've paid far more in other cities, but still nowhere near $1.50 a liter. I guess it helps that I'm using a city-owned bulk distribution network and I live along one of the world's biggest rivers.
If we had the same price discrepancy on gasoline, it'd be cheaper to generate our own electricity from that instead of using the grid. Hell, we'd probably try to figure out safe ways to cook with it, heat with it, and dilute it to use as a cleaning solvent.
They're not really that different. A monopoly, oligopoly, or cartel just artificially limits supply.
Here are different inputs to supply and demand: subsidies artificially increase supply (and therefore often artificially increase demand), monopolies artificially limit supply (and sometimes artificially limit demand due to the high price, in favor of margins over total revenue), embargoes artificially limit supply (which often leads to black and gray markets), price caps artificially increase demand (by keeping the deterrent to purchase out of the way), excises artificially lower demand, tariffs restrict both supply and demand in the importing country and raise supply in the exporting country, production limits artificially reduce supply and sometimes psychologically induce increased demand (think collectibles like Beanie Babies)
Nothing really breaks the laws of supply and demand. There are many ways, though, to finagle them.
In several parts of the US, there is the option to get a carrier that doesn't charge for incomding texts. U.S. Cellular is one of those, at least on most plans.
This depends a great deal on your carrier. U.S. Cellular is mine, and things are a bit different.
I don't pay for incoming calls or incoming texts. I get 900 minutes to anywhere in the US, nights between 2100 and 0900 free, weekends from 2100 Fridays to 0900 Monday free, free calls to everyone on the same network, free calls to my voicemail from anywhere in the US, roaming's included across the US and parts of Canada, and I get unlimited Net data use on the phone. I pay about $71 a month including taxes and fees.
There are downsides, too, though. My unlimited data only works through a WAP proxy, and they want $4.99 more a month to give me access to a real web browser not limited to their WAP proxy with limited content. My data plan is limited to use by the phone for the phone, and isn't for use as a modem by a laptop. It's possible for me to set up my own WAP proxy for the phone, though, and I've played with one as a toy but I found I really only check what's on their limited plan from my phone anyway. Others might want to read whole full-size websites from their phone, but not me. I'd really like if they let me use the phone for my PC's data needs for that $10 part of my plan, but of course that's not going to happen.
The hardware, real estate (although lower than rights-of-way for every hardline) and electricity costs are real. Nearly all the technological advances came from existing landlines, packet radio, and the Internet more so than being invented from scratch.
Hell, CDMA/Edge is very close to being a roaming wireless Ethernet and TDMA/GSM is close to being a roaming wireless Token Ring or FDDI. The AT command set of GSM phones comes from modems. The voice data transmitted across a cell network is encoded with audio codecs just like for landlines, audio files on computers, or VoIP (although often using different codecs specifically designed for cellular telephony).
Every cellphone is basically a cellular radio WAN bridge, a NIC, a CPU/DSP combination encoding and decoding packet data, and a user interface for getting sound, numbers, and text into and out of the system. With sufficient standards, cheap enough spectrum, and enough interoperability testing we could easily have Open Source/Open Hardware phones running commodity hardware and software on commodity networks like chat software on PCs across the Internet.
Spectrum's not cheap, though, and interoperability isn't always easy. Google's play to get the open access rules in the recent US spectrum auction is a big step towards a wireless Internet extravaganza, but only a step.
Laziness of people keeping them from grabbing a real computer and instead opting for holding on to the little phone has a lot to do with it. IMs are zero marginal cost once you have Internet service. SMS usually cost quite a bit.
Just use a laptop (or phone handset) with WiFi and a cable/DSL/fiber/tower-to-customer wireless Internet connection at the other end of that WiFi and you can save a bunch of money.
Sure, in some situations/locations, you get SMS without getting a decent Net connection, but I see people sitting on couches or in coffee shops using SMS.
You could print letters in gold. It's been done for centuries. Getting the gold into a fine powder that's suspended in a solvent that sprays through your inkjet and doesn't void your manufacturer's warranty, now that's a trick.
Well, "going round to my mate's house" for the friend I call the most means driving for 45 minutes or more. I think I'll stick with calling when all I want to say is "what's up?".
The (monetary) efficiency of the final payload is exactly what interests the author. That so much more data is used to send a paltry 160 characters is part of thew reason SMS fairs so poorly from a cost standpoint.
Whether it's avoidable considering Hubble always sends the data to the same few places vs. the needed routing for millions of cell phones to send to millions of other cell phones is a fair question, though.
One way the cost per byte of final payload can be lowered is to allow more bytes of payload per message so the percentage of transmitted data that is overhead shrinks. However, in the case of SMS, we get into the cost factor for the hardware, battery, radio, and software in the handsets to try to change that ratio.
You make a point, but it turns out his project isn't another on-line messaging system. I'm not sure how much competition there is with his project, but it's surely less than IM apps. It's an n-body simulation system.
First, I disagree fully that the suggestion of Mechanical Turk is "Offtopic" for this discussion. Some moderator seems to have gotten trigger happy.
A little bit of super cheap by-the-feature labor might be useful. It might help to recruit someone who's interested in the problem. It might help attract more interest as a project when the updates are flowing from more than one contributor. Replace MT contributors with volunteers as they become available if you like. Take donations from users who don't have sufficient programing skills to pay for the MT work.
I've never used the particular service myself, so I'm not sure what quality you'd get. It's entirely possible, in theory, for that to be a viable option though.
Go to bed, Steve. You'll have a lot of chairs to throw later this week when you realize quality control doesn't test your OS updates on AMD.
I agree, but it's probably better at this point if he can get the domain-aware users ready to program than to get the CS students up to speed on his subject.
Which open source and closed source companies makes a big difference here. Which companies you consider one or the other does, too.
IBM, Sun, Adobe, Corel, Novell, and Apple do a bit of both. Red Hat is considered an open source company, and their books look pretty good. Google uses Open Source software and gives part of their code improvements back even though they're using the code internally. SCO is a closed source company, as are Borland, Fog Creek Software, IMSI, Opera Software, and thousands more. Lernout & Hauspie was a closed-source company, and they are now gone (leaving the Dragon Software founders in the lurch, actually, after buying that company out with stock that ended up unsellable). For every Microsoft, HP, IBM, Sage, or Adobe there are thousands or tens of thousands of little software companies.
One market where OSS really helps is in the really small freelance programmer and software consulting segment. Many of these companies don't have a packaged product to sell, but sell custom programming to multiple clients. They can use OSS as a starting point and customize it rather than building a huge library of software from scratch. This gets smaller software companies lots of business earlier in the operation of their companies, because they can undercut the closed-source competition. The money might be spread out more instead of being in one person's pockets, but most programmers aren't the chairman of Microsoft, and are happy to make a decent living doing what they do.
Since most of the software in the world isn't sold in shrink-wrapped packages, you might want to consider the importance of custom development.