Freezing rain is noticeable worse on elevated roads. Even the raised areas before and after bridges are much worse than the road at ground level. The actual bridges are typically horrible. You need to be careful of bridges in general because they easily swing with air temp and/or sun exposure, causing snow to melt then freeze, even when below freezing temps for several days.
We don't have that issue at all with cold air + warm ground. Cold snaps during the Fall and early Winter are the easiest snow to drive in. I wonder if it has to do with asphalt instead of concrete. Asphalt roads are horribly slick during rain, but I have no issues with concrete, which is what most roads around here are. Asphalt roads are typically associated with cheap back roads to poor neighborhoods.
Snow around here is rarely slick. In order to be slick, it needs to be slushy, which typically requires really cold temps and lots of salt mixed in keeping it in a partially frozen state. Another way is for the ground to be really really cold and the air to be warm, so the snow melts from the air, then refreezes. I'm sure there are other ways, but those are the ones I'm most familiar with.
I was going to say. The person defends their self with the reason for the issues was not having plows and salt trucks. The issue was 2-3 inches causing such chaos. I've done my fair share of driving in 6"-8" of unplowed snow and some times freeezing rain in both cities and on highways
Manufacturing for public sale to make a profit is easy to track, but tracking private use is nearly impossible. This is the new "don't record TV shows with your VCR" type issue.
It's been a bit since I've seen the tech specs, but ARMs CPUs do not scaling well with frequency at all. Something like 900mhz is 2 watts and 1.2ghz is 10watts. ARM is highly optimized for low frequency small cache designs. As its scales up, it becomes worse than modern x64 cpus. Not to say ARM won't improve, but you can't say it's simple.
In which ways? Branchy code? MIPS per joule? Flops per Joule? Raw MIPS or FLOPS? Performance per core? Otherwise you're just saying something like "My semi beat your car".
I find that logic faulty as I must kill plants in order to survive. Heck, I have to kill bacteria. Without killing, I myself would die on a daily basis. The logic should apply at all levels, which it does not. I do value sentient beings in general, so obviously humans, but also the possibility of aliens, computer AIs, and I currently think dolphins and some whales may be smart enough to fall into the category.
I also think that allowing someone to do something is no different than doing it yourself, assuming you are in a position to stop it. A pacifist that is in the position to stop someone from harming another is no better than the person who committed the harm. A person in a position of power must take responsibility for those whom they have power over, even if not an official position of power, but one of circumstance.
I'm not judging anyone, just stating my opinion. I am human and have my own flaws.
Yes, that's the answer to everything. Don't like our government, just leave the USA. Screw friends and family, vote with your wallet! Teach us how to punish ourselves to make these big corps feel bad, Obi-Wan.
1gb with 1,000 ms latency means 125MB of buffer, multiply by 1,000 ports and you're talking about over 100GB of network buffers on a single switch. When you're talking about a modern Ethernet based network, latency indicates congestion which means you don't have 1gb of bandwidth. You don't get latency, you get packet loss.
It doesn't matter, it's cheaper. 1gb fiber is cheaper than POTS, yet alone DSL or Cable, in the same way your Rasberri Pi is cheaper than the ENIAC. You get cheaper and faster all at the same time.
If you're waiting for buffering, blame your ISP. Netflix and YouTube both play near instantly(less than 1sec, closer to 1/2 sec) for me and seeking is just as good, even with 1080p. 4K resolutions do take 2-3 seconds to buffer, but fine once it gets playing, but seeking is annoying. Maybe I should upgrade to 100mb/s.
That's old. We already have 16tb/s per fiber using currently used fiber. The newer fiber tech being tested is already about 850tb/s per fiber. 1.6tb/s is so slow, current research tech has shown transferring 1tb/s over 12,000km of fiber. That's almost 1/2 around the Earth at its widest part.
They didn't amend their agreement in the normal sense, they clarified it. Parts of their ToS already mentioned you could do non-commercial servers for stuff like friends and family, even game servers, but another part of their ToS was written in an absolute tone saying that no servers of any kind were allowed. It was also asked on their official forums and they said personal servers are fine, just don't be making money on it.
Also there is the question of congestion, or rather lack of it. I can't imagine Google is doing point-to-point fiber. It is probably GPON. That means the more subscribers in an area, the less speed.
They use wmd-PON, which gives each end point it's own dedicated 1.25gb/s of bandwidth with no sharing. My understanding is they can handle about 400 customers per network chassis and each chassis can support 400gb of uplink. These chassis plug directly into their core route which has 100gb ports.
I'm not sure what kind of routing speed their core routers have, but with current tech, one could roll out 1gb/1gb to 500,000 houses, and give them non-blocking full speed, even on the uplink. That kind of design would be expensive and there is no reason to assume a large population would need a full non-blocking design at the core, but from the last mile to the core, it is simple to do non-blocking.
Slashdot Mirror
Freezing rain is noticeable worse on elevated roads. Even the raised areas before and after bridges are much worse than the road at ground level. The actual bridges are typically horrible. You need to be careful of bridges in general because they easily swing with air temp and/or sun exposure, causing snow to melt then freeze, even when below freezing temps for several days.
When I grew up in a city of 30,000 people, 45 minutes of notice was enough time to close schools. How would 4 hours not be enough?
We don't have that issue at all with cold air + warm ground. Cold snaps during the Fall and early Winter are the easiest snow to drive in. I wonder if it has to do with asphalt instead of concrete. Asphalt roads are horribly slick during rain, but I have no issues with concrete, which is what most roads around here are. Asphalt roads are typically associated with cheap back roads to poor neighborhoods.
Snow around here is rarely slick. In order to be slick, it needs to be slushy, which typically requires really cold temps and lots of salt mixed in keeping it in a partially frozen state. Another way is for the ground to be really really cold and the air to be warm, so the snow melts from the air, then refreezes. I'm sure there are other ways, but those are the ones I'm most familiar with.
I was going to say. The person defends their self with the reason for the issues was not having plows and salt trucks. The issue was 2-3 inches causing such chaos. I've done my fair share of driving in 6"-8" of unplowed snow and some times freeezing rain in both cities and on highways
Manufacturing for public sale to make a profit is easy to track, but tracking private use is nearly impossible. This is the new "don't record TV shows with your VCR" type issue.
It's been a bit since I've seen the tech specs, but ARMs CPUs do not scaling well with frequency at all. Something like 900mhz is 2 watts and 1.2ghz is 10watts. ARM is highly optimized for low frequency small cache designs. As its scales up, it becomes worse than modern x64 cpus. Not to say ARM won't improve, but you can't say it's simple.
Ahh, yes. Solaris "Hardware threads". Barrel processors, like hyper threading, but with 4-8 virtual threads instead of 2.
High volume is reliable income that can be planned around.
Or define "winning" as having a higher revenue, higher margins, more R&D, and more fab plants, when comparing any single other company to Intel.
IP will be ignored and it will be impossible or impractical to enforce
When I tried to look up "Unenforceable law", I got forwarded to "void" and "invalid" laws. I guess we know what IP law is going to turn into.
In which ways? Branchy code? MIPS per joule? Flops per Joule? Raw MIPS or FLOPS? Performance per core? Otherwise you're just saying something like "My semi beat your car".
And harddrives should not need more than 4KB of cache.
They use the Men in Black Memory Eraser after copying your ideas and destroy all back-ups and copies.
I find that logic faulty as I must kill plants in order to survive. Heck, I have to kill bacteria. Without killing, I myself would die on a daily basis. The logic should apply at all levels, which it does not. I do value sentient beings in general, so obviously humans, but also the possibility of aliens, computer AIs, and I currently think dolphins and some whales may be smart enough to fall into the category.
I also think that allowing someone to do something is no different than doing it yourself, assuming you are in a position to stop it. A pacifist that is in the position to stop someone from harming another is no better than the person who committed the harm. A person in a position of power must take responsibility for those whom they have power over, even if not an official position of power, but one of circumstance.
I'm not judging anyone, just stating my opinion. I am human and have my own flaws.
That is not the ethical position of Christians in general. The Bible says that we should not kill with malice.
It's the only way to survive. It's like saying "It's not right to kill someone even if you life depends on it because they're attacking you"
Yes, that's the answer to everything. Don't like our government, just leave the USA. Screw friends and family, vote with your wallet! Teach us how to punish ourselves to make these big corps feel bad, Obi-Wan.
1gb with 1,000 ms latency means 125MB of buffer, multiply by 1,000 ports and you're talking about over 100GB of network buffers on a single switch. When you're talking about a modern Ethernet based network, latency indicates congestion which means you don't have 1gb of bandwidth. You don't get latency, you get packet loss.
It's just overkill.
It doesn't matter, it's cheaper. 1gb fiber is cheaper than POTS, yet alone DSL or Cable, in the same way your Rasberri Pi is cheaper than the ENIAC. You get cheaper and faster all at the same time.
If you're waiting for buffering, blame your ISP. Netflix and YouTube both play near instantly(less than 1sec, closer to 1/2 sec) for me and seeking is just as good, even with 1080p. 4K resolutions do take 2-3 seconds to buffer, but fine once it gets playing, but seeking is annoying. Maybe I should upgrade to 100mb/s.
To sum it up, Provo gave up millions of dollars a year in revenue
They were bleeding a net loss of $1m/year. They cut their losses while retaining most of the benefit.
That's old. We already have 16tb/s per fiber using currently used fiber. The newer fiber tech being tested is already about 850tb/s per fiber. 1.6tb/s is so slow, current research tech has shown transferring 1tb/s over 12,000km of fiber. That's almost 1/2 around the Earth at its widest part.
They didn't amend their agreement in the normal sense, they clarified it. Parts of their ToS already mentioned you could do non-commercial servers for stuff like friends and family, even game servers, but another part of their ToS was written in an absolute tone saying that no servers of any kind were allowed. It was also asked on their official forums and they said personal servers are fine, just don't be making money on it.
Also there is the question of congestion, or rather lack of it. I can't imagine Google is doing point-to-point fiber. It is probably GPON. That means the more subscribers in an area, the less speed.
They use wmd-PON, which gives each end point it's own dedicated 1.25gb/s of bandwidth with no sharing. My understanding is they can handle about 400 customers per network chassis and each chassis can support 400gb of uplink. These chassis plug directly into their core route which has 100gb ports.
I'm not sure what kind of routing speed their core routers have, but with current tech, one could roll out 1gb/1gb to 500,000 houses, and give them non-blocking full speed, even on the uplink. That kind of design would be expensive and there is no reason to assume a large population would need a full non-blocking design at the core, but from the last mile to the core, it is simple to do non-blocking.