Wow... entire Solaris production systems have failed... I'm sure that's never happened to a w2k installation... the OS protects your computer from hardware failures.
The "leading auction site" comparision is really apples-and-oranges: which is easier to run stable system for: front end web service, or an eBay sized database?
isn't it amusing that Compaq offers 99.99% availiablity guarantees for Tru64 and OpenVMS, but only 99.9% for NT on a page that Microsoft points to?
BTW what's a "secure web site"? Is that one like CDUniverse?
Using an unproven rocket as an example to quantify all rocket failures is at best uneducated. No one is going to launch a Mars mission on an unproven rocket; try PROTON or Shuttle.
In any case, this misses the point, because you need a rocket anyways to get up to your middle ground.
Lemme get this straight: an MD recorder, which costs roughly as much as the Mambo X is crap, because (pick one): 1. Sony makes it, and not some company we've never heard about. 2. It's nonprosecutable to copy music onto MDs (IANAL) 3. It doesn't require a k-rad CD-burner to use 4. It can record live audio and can be used as a mic preamp 5. One player + 3 discs is smaller than a MP3 CD player could possibly be 6. It sounds better than 128k encoded MP3s
If you're willing to make all the above sacrifices, more power to you. I'll stick with my piece of crap minidisc recorder.
Radio channel doesn't know you're listening, but the key here is the aggregate. Advertisers would be dumb to pay without knowing how many people are listening, so there are ways of finding out.
More channels => more variety, yes, but you're still not looking at even 1% of independant talent getting signed, I'll bet.
Listening on the radio still causes money to flow to artists (there are a number of agencies, such as RIAA/ASCAP, that radio stations pay). So, you pay by listening to ads, advertisers pay radio stations, who pay RIAA/ASCAP, who pay artists. MP3s, you pay your ISP for the bandwidth, they pay Sprint/MCI/UUNet, and the money never makes it to the artist.
I disagree about digital/satellite TV leading to broader searches for new talent. BIG RECORD LABELS NEVER SEARCH FOR TALENT. You gotta send them a demo tape; they'll never scope out local clubs. The point: if you do distribution this way, some of the more risk-adverse artists (who may in fact be very good) won't make it into the channel, because they can't sell too many CDs/MP3s.
production cost goes to zero? what studios are you visiting? Studios run around $100/hr (more for better studios). Sure, you don't have to press the CDs, but that's about $1k + 5c/CD, which is nothing compared to the costs of recording, mixing, mastering, marketing,...
So basically you want to say "pirating music is OK, and we can shift the cost off the consumer."
This revenue model is thick with bullshit. First of all, the independant guy (or label) is almost never going to be able to get airtime, so their stuff is ALWAYS going to get stolen, plus they get stuck with the studio costs. Secondly, distribution channel still costs money... bandwidth and all that.
Music is a public good, but I sure don't want to be taxed for it. The current revenue model is the one that I think works best for everybody (don't think that CD is worth $16.98? Then don't buy it; listen to the radio)
Bottom line: I and thousands of other semipros are sinking thousands of dollars into gear and studio time. If everyone pirated music there'd be no possibility of return on investment, so anyone the major record labels passed up won't get distribution.
Disclaimer: I'm taiwanese too, and have been in the states almost all my life. My understanding is from talking to someone from China and working for a global communications company which would like to provide network access in China.
I believe that all internet access in China is filtered through an array of censorship servers (what's worse is users get billed per meg for the priviledge of using these censorship servers).
PEEK and POKE is the way you dink memory directly in BASIC... at least on Commodore 64s. Wow was that a long time ago... I assume that's the reference...
Yes, in short because there are GSM providers (I wish there were more!) in the US. I also like the "non-standard" system here.
While I agree that GSM is better technology, I'd not buy a GSM phone right now (even tho I had one over the summer) because the coverage is too limited. Of course, 2 Mbit might change that, but... But back to the main point, I like the choice... it creates competition and allows providers to find niches. Here (in Pittsburgh) we have AMPS, served by BellAtlantic, which is good for the casual "emergency" and weekends and evenings user. TDMA 900, served by AT&T Wireless Services, which is good for business travelers who travel a lot, especially on the road, and don't want exorbiant analog roaming charges. CDMA 1900, served by Sprint PCS, which is good for a casual traveler who wants to pay less (than AT&T) but still wants roaming to all big cities without exorbiant roaming charges. GSM, served by Ariel, which is good for the European traveler, as well as a city-only casual user (per second rounding is nice...)
Cmpanies don't necessarily put themselves in these niches (ie if I walk up to AT&T Wireless Services and tell them I'm looking for a cheap plan but I want cheap digital roaming, they won't point me towards Sprint PCS), but these are just the niches where they're the cheapest/best plan. If everything were one technology, there'd be no competition.
Also, wireless giants are giants. If there's market, there are people to throw at phones, and there's enough competition that you really do want to innovate. Europe's innovating GSM for us, Sprint is innovating (kinda... wireless web) on CDMA 1900 for us, AT&T should be (I don't know what they're up to) innovating TDMA 900 (Motorola sure is...), and AMPSi s on it's way out. I think it's the best of both worlds: large companies pay for consumer's choice. I'm not saying our technology is better, it's not. It's just that we have more choice.
802.11? Secure? interop means anyone can read your bits... especially since the 802.11 phy layer spec is out there... dunno if the Linux WaveLAN II drivers have promiscuous mode, but I'm sure you can hear the data (otherwise you can't get the DCF to work)
You can always encrypt at the link layer; Lucent's WaveLAN line includes some encryption that they call WEP (wired equivalent privacy).
It's an EMU10k card though, so if they open the source code, people might improve (horrors) the firmware, and they could lose sales of the EMU10k-based Audio Production Studio.
Of course, I'd love to be able to program the DSP on my sound card to do all sorts of strange stuff... maybe it'll be a supported platform on distributed.net =)
Bits is bits. You're getting out the original data... no generational loss. I don't get your point here... it's not like you're making an analog copy (like on to videotape).
I think you'd have little trouble keeping a store open on Sunday, unless you live in a place where the vast majority of the population is churchgoing.
You can't get your mail on Christmas day for the same reason I can't get mine on the Fourth of July. Until recently, I wasn't an American (citizen) and as far as I know my carrier didn't have to be either. But we celebrate it as a national holiday. Even if I didn't believe that the US ever became independant from the UK, I'd still be "forced" to deal with that national holiday.
Non-Christians don't make laws that restrict Christian freedom.
How about teachers who aren't allowed to bring "Christian" things to school (ie crosses) while no one says a word about Haunnaka, or putting up pictures of witches on Halloween for that matter?
because of certain ones who enjoy restricting people's freedom
If you want to see who's freedom is more repressed, consider whether you would have made the same comment about Muslims, or Buddhists.
The ones around here liked to molest young boys.
Anyone who likes to molest young boys should not be in the ministry. Just because you're a priest doesn't mean your close to God; that's about a personal relationship with Christ. In any case, you don't need religion to treat humans like they're humans. But you do need Christ to love all humans as yourself.
If you're buying your CPU for games, I wholeheartedly agree. If the apps you run are all so mass market that they have to work on Intel/AMD/Cyrix, then you're golden going with the fastest stable CPU.
On the other hand, if you want to do any type of niche work using closed source software, Intel seems to be a better bet. While AMD hasn't had a publicity flop like Intel did over FDIV, I would guess that many developers don't test their code on AMDs, so there's a higher probability of bad bugs.
Anyways, the real thing I'm happy about with cumine is that now we mobile users can have desktop speeds, and not on a mac.
Maybe I'm not in the majority, but there are people who use CPUs for more than compiling kernels (not that I'm flaming people who compile kernels, as I do so myself quite frequently)
Ever run a batch job that runs for 30-40 hours, and have to run 5 more after that's finished? Ever wish you could have realtime effects on multiple channels, downmix and realtime mp3 encode the result at multiple bitrates while streaming an uncompressed version to disk? Sure would save a lot of time in music production. Ever have to wait for photoshop to run a filter?
Sure, CPUs today are probably 90%+ idle, but we can afford to overprovision because human time is so expensive. I wonder what effect CPU speed has on TCO and total marginal benefit of ownership...
First off, I'm primarily a computer scientist... music is a [serious] hobby.
When we talk about computers _replacing_ musicians, I don't think that will happen. Even if a computer can write better music than I ever can, I'll still compose, because it helps me respond to emotion. In fact, many people will likely do the same. With reasonable probability, some of these types of music will be emotionally more satisfying than computer generated score and computer generated lyrics sung and played by computers. After all, if this technology was invented when only classical music was played, pop music would probably still have evolved (maybe differently, but that's the subject for another post) because of the change in culture and expression.
Saying that computers will ever kill music is like saying they'll kill painting or what-not. Technology can simulate some instance of a human's creative spark, but IMO it will never be able to simulate the spark itself. Perhaps this will make it difficult to make a living as a professional recording/performing artist (not that it isn't difficult even now), but such is the nature of technology.
I'd argue that this won't make us less cultured, because people will still pursue music for personal fulfilment; it will just change how people are compensated for music. The change would be similar to how if all software was open-source by law, some people would still program, and in fact the people who stick around will be among the best.
palmtops aren't supposed to be PCs. Until we put enough processing power into one of these things to get good speech recognition, the input capability is too limited. And sure, we have speech recognition today, but the processing power and energy consumption requirements are too great, though IBM seems to have gotten around that problem in a wearable.
natd is maintained... as mentioned before, it executes in user space, so it's easy to hack up yourself. (basically it's a front-end to libalias) active mode ftp (outbound) works for me, and has since 2.2.7 (the first kernel version Then again, sometimes ICQ works without my ugly port configuration, so maybe I'm just lucky.
You know, they have a prototype...
on
Smart Dust
·
· Score: 1
They have a prototype, and they discussed it at Mobicom 99 (in Seattle)... basically they're targeting 1mm^3 but they've got it down to 2cm^3, so it's only a matter of time.
They do seem rather power-hungry though... especially if they try to communicate amongst each other (as opposed to a base station). And come on, if you saw someone coming through your house with a huge laser pickup and dark glasses on, I don't think you'd seriously consider your actions private.
As people have mentioned, the Operations and Maintanence on this constellation is going to cost quite a bit (to say the least). However, if you think about it, this is a system that the Department of Defense may want, and they'll get it for a song. Whether or not you like that...
Phones not working well indoors is by design.
Sure MOT got most of the Iridium money, but much of it was money spent buying parts or doing development, so all they get is a bunch of legacy code. And the experience of negotiating a few launch contracts.
2 year lifetime is something that I don't really buy; they're probably saying that to make sure they don't get sued if their satellites went down too fast. Of course, there are lots of things that can fail on satellites, so you can never be sure.
Those of you who hope that Teledesic will be nice, fast, sanely priced, and well designed, remember who's the prime contractor =)
1. Where's the source? =) 2. If you can't even look at crashdumps and disassemble (ie if you don't have a copy of win2k), isn't it rather hard to execute arbitrary code?
Agreed. Isn't it great that we're getting rid of all the legacy hardware like ISA and Super I/O (which free OSes support fairly well) and replacing them with hip plug-n-pray USB devices? What a coincidence that the central peripheral connection in EasyPC isn't well supported by Linux/*BSD.
Images still get a referrer header, so if the referrer is not from your domain, you send an image that says "please go directly to www.NotEnoughBandwidth.com" People would stop linking to your images really fast.
2400bps? I think your networking background could use some work. Back of the envelope calculations: Let's assume MTU=1500 bytes; no IP/TCP options; unlimited down, restricted up.
14k up means 1.2 full sized packets/s (14k/(8*(1500-40))); or 45 ack-sized packets up.
TCP steady state is 1 ack / 2 pkts, hence 90 pkts/s down (1 Mbps). I'm feeling sorry for you now.
In addition, this is assuming a zero-loss case. If you're transfering over anything lossy, your pkt losses will hurt you more on the data path than the reverse path (retransmits cost more, acks are redundant).
If you want to get really good measurements, see how much ping throughput you get to the border router with packet size of 40 (20 payload, 20 IP).
I'm on DSL, and accept the limited uplink (640k down, 90k up). Yeah it sucks for sending files on the reverse path, but I'd rather that than 365k symmetric.
Slashdot Mirror
Wow... entire Solaris production systems have failed... I'm sure that's never happened to a w2k installation... the OS protects your computer from hardware failures.
The "leading auction site" comparision is really apples-and-oranges: which is easier to run stable system for: front end web service, or an eBay sized database?
isn't it amusing that Compaq offers 99.99% availiablity guarantees for Tru64 and OpenVMS, but only 99.9% for NT on a page that Microsoft points to?
BTW what's a "secure web site"? Is that one like CDUniverse?
Using an unproven rocket as an example to quantify all rocket failures is at best uneducated. No one is going to launch a Mars mission on an unproven rocket; try PROTON or Shuttle.
In any case, this misses the point, because you need a rocket anyways to get up to your middle ground.
Lemme get this straight: an MD recorder, which costs roughly as much as the Mambo X is crap, because (pick one):
1. Sony makes it, and not some company we've never heard about.
2. It's nonprosecutable to copy music onto MDs (IANAL)
3. It doesn't require a k-rad CD-burner to use
4. It can record live audio and can be used as a mic preamp
5. One player + 3 discs is smaller than a MP3 CD player could possibly be
6. It sounds better than 128k encoded MP3s
If you're willing to make all the above sacrifices, more power to you. I'll stick with my piece of crap minidisc recorder.
Radio channel doesn't know you're listening, but the key here is the aggregate. Advertisers would be dumb to pay without knowing how many people are listening, so there are ways of finding out.
More channels => more variety, yes, but you're still not looking at even 1% of independant talent getting signed, I'll bet.
Listening on the radio still causes money to flow to artists (there are a number of agencies, such as RIAA/ASCAP, that radio stations pay). So, you pay by listening to ads, advertisers pay radio stations, who pay RIAA/ASCAP, who pay artists. MP3s, you pay your ISP for the bandwidth, they pay Sprint/MCI/UUNet, and the money never makes it to the artist.
I disagree about digital/satellite TV leading to broader searches for new talent. BIG RECORD LABELS NEVER SEARCH FOR TALENT. You gotta send them a demo tape; they'll never scope out local clubs. The point: if you do distribution this way, some of the more risk-adverse artists (who may in fact be very good) won't make it into the channel, because they can't sell too many CDs/MP3s.
production cost goes to zero? what studios are you visiting? Studios run around $100/hr (more for better studios). Sure, you don't have to press the CDs, but that's about $1k + 5c/CD, which is nothing compared to the costs of recording, mixing, mastering, marketing, ...
So basically you want to say "pirating music is OK, and we can shift the cost off the consumer."
This revenue model is thick with bullshit. First of all, the independant guy (or label) is almost never going to be able to get airtime, so their stuff is ALWAYS going to get stolen, plus they get stuck with the studio costs. Secondly, distribution channel still costs money... bandwidth and all that.
Music is a public good, but I sure don't want to be taxed for it. The current revenue model is the one that I think works best for everybody (don't think that CD is worth $16.98? Then don't buy it; listen to the radio)
Bottom line: I and thousands of other semipros are sinking thousands of dollars into gear and studio time. If everyone pirated music there'd be no possibility of return on investment, so anyone the major record labels passed up won't get distribution.
Disclaimer: I'm taiwanese too, and have been in the states almost all my life. My understanding is from talking to someone from China and working for a global communications company which would like to provide network access in China.
I believe that all internet access in China is filtered through an array of censorship servers (what's worse is users get billed per meg for the priviledge of using these censorship servers).
PEEK and POKE is the way you dink memory directly in BASIC... at least on Commodore 64s. Wow was that a long time ago... I assume that's the reference...
Why not use a Crypto iButton for challenge/response?
www.ibutton.com
Yes, in short because there are GSM providers (I wish there were more!) in the US. I also like the "non-standard" system here.
While I agree that GSM is better technology, I'd not buy a GSM phone right now (even tho I had one over the summer) because the coverage is too limited. Of course, 2 Mbit might change that, but... But back to the main point, I like the choice... it creates competition and allows providers to find niches. Here (in Pittsburgh) we have
AMPS, served by BellAtlantic, which is good for the casual "emergency" and weekends and evenings user.
TDMA 900, served by AT&T Wireless Services, which is good for business travelers who travel a lot, especially on the road, and don't want exorbiant analog roaming charges.
CDMA 1900, served by Sprint PCS, which is good for a casual traveler who wants to pay less (than AT&T) but still wants roaming to all big cities without exorbiant roaming charges.
GSM, served by Ariel, which is good for the European traveler, as well as a city-only casual user (per second rounding is nice...)
Cmpanies don't necessarily put themselves in these niches (ie if I walk up to AT&T Wireless Services and tell them I'm looking for a cheap plan but I want cheap digital roaming, they won't point me towards Sprint PCS), but these are just the niches where they're the cheapest/best plan. If everything were one technology, there'd be no competition.
Also, wireless giants are giants. If there's market, there are people to throw at phones, and there's enough competition that you really do want to innovate. Europe's innovating GSM for us, Sprint is innovating (kinda... wireless web) on CDMA 1900 for us, AT&T should be (I don't know what they're up to) innovating TDMA 900 (Motorola sure is...), and AMPSi s on it's way out. I think it's the best of both worlds: large companies pay for consumer's choice. I'm not saying our technology is better, it's not. It's just that we have more choice.
802.11? Secure? interop means anyone can read your bits... especially since the 802.11 phy layer spec is out there... dunno if the Linux WaveLAN II drivers have promiscuous mode, but I'm sure you can hear the data (otherwise you can't get the DCF to work)
You can always encrypt at the link layer; Lucent's WaveLAN line includes some encryption that they call WEP (wired equivalent privacy).
It's an EMU10k card though, so if they open the source code, people might improve (horrors) the firmware, and they could lose sales of the EMU10k-based Audio Production Studio.
Of course, I'd love to be able to program the DSP on my sound card to do all sorts of strange stuff... maybe it'll be a supported platform on distributed.net =)
Bits is bits. You're getting out the original data... no generational loss. I don't get your point here... it's not like you're making an analog copy (like on to videotape).
I think you'd have little trouble keeping a store open on Sunday, unless you live in a place where the vast majority of the population is churchgoing.
You can't get your mail on Christmas day for the same reason I can't get mine on the Fourth of July. Until recently, I wasn't an American (citizen) and as far as I know my carrier didn't have to be either. But we celebrate it as a national holiday. Even if I didn't believe that the US ever became independant from the UK, I'd still be "forced" to deal with that national holiday.
Non-Christians don't make laws that restrict Christian freedom.
How about teachers who aren't allowed to bring "Christian" things to school (ie crosses) while no one says a word about Haunnaka, or putting up pictures of witches on Halloween for that matter?
because of certain ones who enjoy restricting people's freedom
If you want to see who's freedom is more repressed, consider whether you would have made the same comment about Muslims, or Buddhists.
The ones around here liked to molest young boys.
Anyone who likes to molest young boys should not be in the ministry. Just because you're a priest doesn't mean your close to God; that's about a personal relationship with Christ. In any case, you don't need religion to treat humans like they're humans. But you do need Christ to love all humans as yourself.
If you're buying your CPU for games, I wholeheartedly agree. If the apps you run are all so mass market that they have to work on Intel/AMD/Cyrix, then you're golden going with the fastest stable CPU.
On the other hand, if you want to do any type of niche work using closed source software, Intel seems to be a better bet. While AMD hasn't had a publicity flop like Intel did over FDIV, I would guess that many developers don't test their code on AMDs, so there's a higher probability of bad bugs.
Anyways, the real thing I'm happy about with cumine is that now we mobile users can have desktop speeds, and not on a mac.
Maybe I'm not in the majority, but there are people who use CPUs for more than compiling kernels (not that I'm flaming people who compile kernels, as I do so myself quite frequently)
Ever run a batch job that runs for 30-40 hours, and have to run 5 more after that's finished? Ever wish you could have realtime effects on multiple channels, downmix and realtime mp3 encode the result at multiple bitrates while streaming an uncompressed version to disk? Sure would save a lot of time in music production. Ever have to wait for photoshop to run a filter?
Sure, CPUs today are probably 90%+ idle, but we can afford to overprovision because human time is so expensive. I wonder what effect CPU speed has on TCO and total marginal benefit of ownership...
First off, I'm primarily a computer scientist... music is a [serious] hobby.
When we talk about computers _replacing_ musicians, I don't think that will happen. Even if a computer can write better music than I ever can, I'll still compose, because it helps me respond to emotion. In fact, many people will likely do the same. With reasonable probability, some of these types of music will be emotionally more satisfying than computer generated score and computer generated lyrics sung and played by computers. After all, if this technology was invented when only classical music was played, pop music would probably still have evolved (maybe differently, but that's the subject for another post) because of the change in culture and expression.
Saying that computers will ever kill music is like saying they'll kill painting or what-not. Technology can simulate some instance of a human's creative spark, but IMO it will never be able to simulate the spark itself. Perhaps this will make it difficult to make a living as a professional recording/performing artist (not that it isn't difficult even now), but such is the nature of technology.
I'd argue that this won't make us less cultured, because people will still pursue music for personal fulfilment; it will just change how people are compensated for music. The change would be similar to how if all software was open-source by law, some people would still program, and in fact the people who stick around will be among the best.
palmtops aren't supposed to be PCs. Until we put enough processing power into one of these things to get good speech recognition, the input capability is too limited. And sure, we have speech recognition today, but the processing power and energy consumption requirements are too great, though IBM seems to have gotten around that problem in a wearable.
natd is maintained... as mentioned before, it executes in user space, so it's easy to hack up yourself. (basically it's a front-end to libalias) active mode ftp (outbound) works for me, and has since 2.2.7 (the first kernel version Then again, sometimes ICQ works without my ugly port configuration, so maybe I'm just lucky.
They have a prototype, and they discussed it at Mobicom 99 (in Seattle)... basically they're targeting 1mm^3 but they've got it down to 2cm^3, so it's only a matter of time.
They do seem rather power-hungry though... especially if they try to communicate amongst each other (as opposed to a base station). And come on, if you saw someone coming through your house with a huge laser pickup and dark glasses on, I don't think you'd seriously consider your actions private.
As people have mentioned, the Operations and Maintanence on this constellation is going to cost quite a bit (to say the least). However, if you think about it, this is a system that the Department of Defense may want, and they'll get it for a song. Whether or not you like that...
Phones not working well indoors is by design.
Sure MOT got most of the Iridium money, but much of it was money spent buying parts or doing development, so all they get is a bunch of legacy code. And the experience of negotiating a few launch contracts.
2 year lifetime is something that I don't really buy; they're probably saying that to make sure they don't get sued if their satellites went down too fast. Of course, there are lots of things that can fail on satellites, so you can never be sure.
Those of you who hope that Teledesic will be nice, fast, sanely priced, and well designed, remember who's the prime contractor =)
1. Where's the source? =)
2. If you can't even look at crashdumps and disassemble (ie if you don't have a copy of win2k), isn't it rather hard to execute arbitrary code?
Agreed. Isn't it great that we're getting rid of all the legacy hardware like ISA and Super I/O (which free OSes support fairly well) and replacing them with hip plug-n-pray USB devices? What a coincidence that the central peripheral connection in EasyPC isn't well supported by Linux/*BSD.
Images still get a referrer header, so if the referrer is not from your domain, you send an image that says "please go directly to www.NotEnoughBandwidth.com" People would stop linking to your images really fast.
2400bps? I think your networking background could use some work. Back of the envelope calculations: Let's assume MTU=1500 bytes; no IP/TCP options; unlimited down, restricted up.
14k up means 1.2 full sized packets/s (14k/(8*(1500-40))); or 45 ack-sized packets up.
TCP steady state is 1 ack / 2 pkts, hence 90 pkts/s down (1 Mbps). I'm feeling sorry for you now.
In addition, this is assuming a zero-loss case. If you're transfering over anything lossy, your pkt losses will hurt you more on the data path than the reverse path (retransmits cost more, acks are redundant).
If you want to get really good measurements, see how much ping throughput you get to the border router with packet size of 40 (20 payload, 20 IP).
I'm on DSL, and accept the limited uplink (640k down, 90k up). Yeah it sucks for sending files on the reverse path, but I'd rather that than 365k symmetric.