In the case of the BitTorrent Sandvine filtering incident, Wireshark logs could be taken at both ends of a connection (sync the captures over the phone or whatever).
Compare the logs - If RST packets are detected coming in at one end of the connection that were never sent at the other end, that's proof that someone (in this case the ISP) injected them into the connections to shut them down prematurely.
In addition to that, the oven could be modified to either be fully heated or at least preheated by a solar concentrator.
Solar thermal is a LOT cheaper and easier than solar photovoltaic. The problem is that concentrator-based designs can't work in clouds, while PV and nonconcentrated can. Nonconcentrated thermal doesn't work well for electrical energy generation. (Great for hot water heating though.)
The high directionality of antennas such as the DB8 (or CM 4228) should reduce multipath significantly to begin with, in addition to increasing signal strength.
And yes, there is also that issue of preamp overloading where a preamp can make things worse. WNJB in New Brunswick, NJ made it difficult if not impossible to receive NYC HDTV signals at my parents' house a few miles north of NB - Even though WNJB was not in the antenna's main lobe (pretty much at 90 degrees), it had a tendency to overload any preamp.
A CM 4228 or similar would've fixed this as it had much more directionality. They were (and still are) using a normal V/U antenna that they refuse to move from the attic onto the roof.
They'll pretty much have to go to the roof or get cable next year though...
Yeah, I saw the info on that, it appears pretty similar to the CM 4228, pretty hard to beat that one.
The submitter has pretty much three options in order he should try: 1) Get the antenna outside on the porch somehow 2) Get a mastmount preamp 3) Give up and get cable or satellite
That DB-8 antenna he has is pretty similar to the Channel Master 4228, which is one of the highest gain UHF antennas you can find.
The guy that submitted this needs to: 1) Move that antenna to the porch! 2) Try a mast-mount preamp (warning, if one of those stations is nearby and very strong this can make things worse.) 3) Get cable
While Winegard is a pretty good brand, that looks like a crappy antenna. I'd suggest something from the Channel Master 422x series (or Winegard's equivalent), unless some of those HD channels are VHF, in which case it gets MUCH harder to find a compact high gain antenna.
Nearly all of the B&M electronics retailers sell absolutely horribly shitty antennas. (There are occasionally decent ones but it's RARE.)
If you want to get a good antenna you need to go to a specialty store (likely online) or in many cases you'll have luck at home improvement stores like Home Depot or Lowes.
Look for products from Channel Master or Winegard. Both make good antennas and preamps. There are a few other good brands but those are the two that come to mind first.
If you fail with CM or Winegard - get cable. Unfortunately reliable terrestrial HD can be difficult. I don't even bother in my apartment. Everything else about your setup is fine, your OS makes no difference if reception is bad. Garbage in, garbage out.
At Canadian Wal-Marts near Banff and Jasper National Parks, there were quite a few RVs in the parking lot last time I was there (Early 2000s), and I have seen RVs quite a few times in the Wal-Mart parking lot in Vestal, NY.
From another article: "Unless managers decide to yank the welcome mat, or local ordinances prohibit overnight parking, the lots are open to RVs." - So your town could have a local law against overnight parking.
Last I heard (and I have seen evidence of this) is that Wal-Mart effectively has an official company policy that RVs and similar vehicles are allowed to park in their parking lots overnight, even for extended periods. Most other businesses would call the cops or chase the RVs away.
The rationale for Wal-Mart? The people in that camper parked in the parking lot are likely going to go for the most convenient supply shopping available.
1) To get a broad DoF, you need a smaller physical aperture, which requires a HIGHER F-number. It is large-aperture lenses (low F-number) that are difficult to achieve with large film/sensor sizes. Small-aperture is easy - this is why MF lenses are almost always slower than 35mm lenses (you rarely if ever see f/2.8 MF lenses, while f/2.8 35mm primes are common)
2) Reread my previous comment. Read the linked Wikipedia article on the Scheimpflug Principle. The plane of focus doesn't have to be parallel to the film plane - Tilt lenses (and in the case of LF, movements) allow you to make the plane of focus angled with comparison to film. Adams made extremely heavy use of this. Small apertures were one of the reasons he was able to get such deep depth of field, but the other reason is that the lens would be tilted downwards so that the plane of focus was angled instead of straight up and down.
I was mainly addressing that article's comment about people having trouble duplicating Adams' work - That's not because of lack of advance in imaging technology, it's because some mechanical features Adams used heavily are still extremely rare and expensive.
"Why is it that with over 60 years of improvements in cameras, lens sharpness and film grain, resolution and dynamic range that no one has been able to equal what Ansel Adams did back in the 1940s?" Ansel Adams used a large format system. While few people shoot LF anymore, LF is still unbeatable by any other technology. Same idea as the whole "larger sensor is better" in digital. The problem is that if anyone could make a sensor that large, it would cost millions of dollars per unit.
Also, LF gives you the flexibility of (rare for 35mm and MF) tilt/shift lenses, plus some additional tricks that T/S lenses don't give you. Adams relied HEAVILY on some of these tricks (called "movements" in LF) to get his images.
Usenet - there's nothing else that requires splitting files into small chunks. FTP can easily handle 500+ megabyte files, so can pretty much any protocol except for NNTP. (NNTP can probably handle it too, but the news servers likely all impose attachment size limits, and NNTP may not support resuming failed connections.)
So the torrents do it because the release groups do it. The release groups do it because of Usenet.
Thus the torrents are doing it indirectly due to Usenet.
Somehow my previous post on this subject failed to actually get posted.
When it comes to the binaries groups, there is no place for nostalgia. People clinging on to Usenet is proving to be a major pain in the ass for BitTorrent users, and is extremely detrimental to many torrents.
Why is that? It's simple - Thanks to alt.binaries.*, a lot of Torrents are STILL distributed as a bunch of small RARs. This is counterproductive for two reasons: 1) Most of the content was compressed to begin with and trying to compress it again will frequently result in a slight INCREASE in file size 2) Since the seedable form of the content (RARs) and the usable form (the actual content) are not the same, it takes twice the hard drive space to both seed and use the content. End result - People stop seeding much sooner than if the torrent was provided in readily usable form because they need to clear up hard drive space.
We have yet to achieve nuclear fusion that can "break even" and produce more energy than it consumes.
But we have achieved relatively simple devices that do a very good job of generating neutrons (such as the Farnsworth Fusor). They operate at a net loss - But what if you use such a device to bombard fissionable material with neutrons? The idea is similar - The fissionable material would normally be sub-critical, you would effectively "turn it on" by turning on the fusion device.
Talk to your printer manufacturer. (Let me guess - Lexmark?)
More specifically, don't buy from them again. Buy from someone that documents their printer protocols and even produces drivers like Epson. (Although I must admit I'm disappointed in the current state of Epson scanners - while my printer works great, the scanner portion is limited to 8 bits/channel by Epson's closed source drivers.)
Some hardware manufacturers produce good drivers for their products. (NVidia) Some provide documentation so that drivers can be written. (Many) Some provide both documentation and open-source drivers. (Ralink, for example. Intel too.) Some provide neither. You should not give these people your money.
Also, FYI, most if not all printer drivers are implemented in userspace, so your comment isn't really appropriate to a kernel discussion.
H3DII-50 isn't out yet, only the H3DII-39 at this point. IIRC it was announced around the same time as the Kodak sensor, which I think the Hassy uses.
Actually I just looked it up - http://www.hasselbladusa.com/promotions/50-promotion.aspx - the 50 is not available until October 2008 and DOES use the new Kodak sensor, the promotion is that if you buy a 39 before September 30, you get a discount on a lens and an opportunity to upgrade to the 50 for the difference in list price.
Well it would probably be a bad idea to do this pre-emptively prior to an HIV infection, but once someone was infected - If it's a choice between "dead T-cells/no T-cells" and "potentially malfunctioning T-cells", "potentially malfunctioning" is better than "none/dead".
Yup, that's what many VPNs use. IPSEC uses either tunneling through UDP or a completely different protocol that is neither TCP nor UDP (but is layered over IP). I forget what it is called.
The problem with an application like BT using UDP at the application level (effectively developing their own transport protocol) is that it is likely to be more aggressive and less network-friendly than the existing TCP-based schemes. The end result is that by throttling TCP connections, network providers are shooting themselves in a foot - the only possible end result is that they will create a monster. (With that "monster" being an encrypted transport protocol that is far less "nice" in terms of congestion control than TCP.)
Re:NAS: Western Digital MyBook World Edition II
on
What NAS To Buy?
·
· Score: 2, Informative
That's interesting to hear that you think the performance is good - The ratings of that drive on NewEgg are abysmal mainly due to performance issues, average ratings of 2-3 eggs!
Exactly. If the backlog is in the router, the modem's buffer stays empty and latency is low.
If the backlog is in the modem, packet loss does not skyrocket much but latency does. This is a sign of a large buffer. If the buffer were small as you claim, there would be a packet loss increase but no latency increase.
This is not true. Linksys switched to VxWorks not because of lawsuits (otherwise why continue with the GL?), but because it was perceived to have a lower memory footprint than Linux, with the license cost being less than the cost of the "extra" hardware needed to run Linux. Of course, the DD-WRT crew proved them wrong after a few months by trimming down their distro to fit in the reduced-memory devices while retaining as much functionality as the stock firmware.
Slashdot Mirror
In the case of the BitTorrent Sandvine filtering incident, Wireshark logs could be taken at both ends of a connection (sync the captures over the phone or whatever).
Compare the logs - If RST packets are detected coming in at one end of the connection that were never sent at the other end, that's proof that someone (in this case the ISP) injected them into the connections to shut them down prematurely.
In addition to that, the oven could be modified to either be fully heated or at least preheated by a solar concentrator.
Solar thermal is a LOT cheaper and easier than solar photovoltaic. The problem is that concentrator-based designs can't work in clouds, while PV and nonconcentrated can. Nonconcentrated thermal doesn't work well for electrical energy generation. (Great for hot water heating though.)
The high directionality of antennas such as the DB8 (or CM 4228) should reduce multipath significantly to begin with, in addition to increasing signal strength.
And yes, there is also that issue of preamp overloading where a preamp can make things worse. WNJB in New Brunswick, NJ made it difficult if not impossible to receive NYC HDTV signals at my parents' house a few miles north of NB - Even though WNJB was not in the antenna's main lobe (pretty much at 90 degrees), it had a tendency to overload any preamp.
A CM 4228 or similar would've fixed this as it had much more directionality. They were (and still are) using a normal V/U antenna that they refuse to move from the attic onto the roof.
They'll pretty much have to go to the roof or get cable next year though...
Yeah, I saw the info on that, it appears pretty similar to the CM 4228, pretty hard to beat that one.
The submitter has pretty much three options in order he should try:
1) Get the antenna outside on the porch somehow
2) Get a mastmount preamp
3) Give up and get cable or satellite
That DB-8 antenna he has is pretty similar to the Channel Master 4228, which is one of the highest gain UHF antennas you can find.
The guy that submitted this needs to:
1) Move that antenna to the porch!
2) Try a mast-mount preamp (warning, if one of those stations is nearby and very strong this can make things worse.)
3) Get cable
While Winegard is a pretty good brand, that looks like a crappy antenna. I'd suggest something from the Channel Master 422x series (or Winegard's equivalent), unless some of those HD channels are VHF, in which case it gets MUCH harder to find a compact high gain antenna.
Nearly all of the B&M electronics retailers sell absolutely horribly shitty antennas. (There are occasionally decent ones but it's RARE.)
If you want to get a good antenna you need to go to a specialty store (likely online) or in many cases you'll have luck at home improvement stores like Home Depot or Lowes.
Look for products from Channel Master or Winegard. Both make good antennas and preamps. There are a few other good brands but those are the two that come to mind first.
If you fail with CM or Winegard - get cable. Unfortunately reliable terrestrial HD can be difficult. I don't even bother in my apartment. Everything else about your setup is fine, your OS makes no difference if reception is bad. Garbage in, garbage out.
Hmm. Interesting.
At Canadian Wal-Marts near Banff and Jasper National Parks, there were quite a few RVs in the parking lot last time I was there (Early 2000s), and I have seen RVs quite a few times in the Wal-Mart parking lot in Vestal, NY.
http://www.freecampgrounds.com/othercamps.html says that Wal-Mart corporate encourages RVers to report such signage to them, as it directly conflicts with company policy.
From another article: "Unless managers decide to yank the welcome mat, or local ordinances prohibit overnight parking, the lots are open to RVs." - So your town could have a local law against overnight parking.
At least for a while, everything I read about the successors to the 48 series like my 48GX were inferior to the older units.
How is the 50G? Can it compare to the old 48GX? Or is it flimsy and unreliable like some of the 48's immediate successors?
That's an interesting question.
Last I heard (and I have seen evidence of this) is that Wal-Mart effectively has an official company policy that RVs and similar vehicles are allowed to park in their parking lots overnight, even for extended periods. Most other businesses would call the cops or chase the RVs away.
The rationale for Wal-Mart? The people in that camper parked in the parking lot are likely going to go for the most convenient supply shopping available.
1) To get a broad DoF, you need a smaller physical aperture, which requires a HIGHER F-number. It is large-aperture lenses (low F-number) that are difficult to achieve with large film/sensor sizes. Small-aperture is easy - this is why MF lenses are almost always slower than 35mm lenses (you rarely if ever see f/2.8 MF lenses, while f/2.8 35mm primes are common)
2) Reread my previous comment. Read the linked Wikipedia article on the Scheimpflug Principle. The plane of focus doesn't have to be parallel to the film plane - Tilt lenses (and in the case of LF, movements) allow you to make the plane of focus angled with comparison to film. Adams made extremely heavy use of this. Small apertures were one of the reasons he was able to get such deep depth of field, but the other reason is that the lens would be tilted downwards so that the plane of focus was angled instead of straight up and down.
I was mainly addressing that article's comment about people having trouble duplicating Adams' work - That's not because of lack of advance in imaging technology, it's because some mechanical features Adams used heavily are still extremely rare and expensive.
Interesting, but has some serious flaws.
"Why is it that with over 60 years of improvements in cameras, lens sharpness and film grain, resolution and dynamic range that no one has been able to equal what Ansel Adams did back in the 1940s?"
Ansel Adams used a large format system. While few people shoot LF anymore, LF is still unbeatable by any other technology. Same idea as the whole "larger sensor is better" in digital. The problem is that if anyone could make a sensor that large, it would cost millions of dollars per unit.
Also, LF gives you the flexibility of (rare for 35mm and MF) tilt/shift lenses, plus some additional tricks that T/S lenses don't give you. Adams relied HEAVILY on some of these tricks (called "movements" in LF) to get his images.
http://en.wikipedia.org/wiki/Scheimpflug_principle
http://en.wikipedia.org/wiki/Photocatalysis
http://en.wikipedia.org/wiki/Titanium_dioxide
TiO2 can catalyze a wide variety of chemical reactions in the presence of ultraviolet light (such as sunlight).
Why do the groups in "The Scene" do this?
Usenet - there's nothing else that requires splitting files into small chunks. FTP can easily handle 500+ megabyte files, so can pretty much any protocol except for NNTP. (NNTP can probably handle it too, but the news servers likely all impose attachment size limits, and NNTP may not support resuming failed connections.)
So the torrents do it because the release groups do it.
The release groups do it because of Usenet.
Thus the torrents are doing it indirectly due to Usenet.
Somehow my previous post on this subject failed to actually get posted.
When it comes to the binaries groups, there is no place for nostalgia. People clinging on to Usenet is proving to be a major pain in the ass for BitTorrent users, and is extremely detrimental to many torrents.
Why is that? It's simple - Thanks to alt.binaries.*, a lot of Torrents are STILL distributed as a bunch of small RARs. This is counterproductive for two reasons:
1) Most of the content was compressed to begin with and trying to compress it again will frequently result in a slight INCREASE in file size
2) Since the seedable form of the content (RARs) and the usable form (the actual content) are not the same, it takes twice the hard drive space to both seed and use the content. End result - People stop seeding much sooner than if the torrent was provided in readily usable form because they need to clear up hard drive space.
What's so impressive about his ID number?
A nice idea until a rocket explodes or crashes.
This leads to something I've always wondered.
We have yet to achieve nuclear fusion that can "break even" and produce more energy than it consumes.
But we have achieved relatively simple devices that do a very good job of generating neutrons (such as the Farnsworth Fusor). They operate at a net loss - But what if you use such a device to bombard fissionable material with neutrons? The idea is similar - The fissionable material would normally be sub-critical, you would effectively "turn it on" by turning on the fusion device.
Talk to your printer manufacturer. (Let me guess - Lexmark?)
More specifically, don't buy from them again. Buy from someone that documents their printer protocols and even produces drivers like Epson. (Although I must admit I'm disappointed in the current state of Epson scanners - while my printer works great, the scanner portion is limited to 8 bits/channel by Epson's closed source drivers.)
Some hardware manufacturers produce good drivers for their products. (NVidia)
Some provide documentation so that drivers can be written. (Many)
Some provide both documentation and open-source drivers. (Ralink, for example. Intel too.)
Some provide neither. You should not give these people your money.
Also, FYI, most if not all printer drivers are implemented in userspace, so your comment isn't really appropriate to a kernel discussion.
H3DII-50 isn't out yet, only the H3DII-39 at this point. IIRC it was announced around the same time as the Kodak sensor, which I think the Hassy uses.
Actually I just looked it up - http://www.hasselbladusa.com/promotions/50-promotion.aspx - the 50 is not available until October 2008 and DOES use the new Kodak sensor, the promotion is that if you buy a 39 before September 30, you get a discount on a lens and an opportunity to upgrade to the 50 for the difference in list price.
Well it would probably be a bad idea to do this pre-emptively prior to an HIV infection, but once someone was infected - If it's a choice between "dead T-cells/no T-cells" and "potentially malfunctioning T-cells", "potentially malfunctioning" is better than "none/dead".
Yup, that's what many VPNs use. IPSEC uses either tunneling through UDP or a completely different protocol that is neither TCP nor UDP (but is layered over IP). I forget what it is called.
The problem with an application like BT using UDP at the application level (effectively developing their own transport protocol) is that it is likely to be more aggressive and less network-friendly than the existing TCP-based schemes. The end result is that by throttling TCP connections, network providers are shooting themselves in a foot - the only possible end result is that they will create a monster. (With that "monster" being an encrypted transport protocol that is far less "nice" in terms of congestion control than TCP.)
That's interesting to hear that you think the performance is good - The ratings of that drive on NewEgg are abysmal mainly due to performance issues, average ratings of 2-3 eggs!
Exactly. If the backlog is in the router, the modem's buffer stays empty and latency is low.
If the backlog is in the modem, packet loss does not skyrocket much but latency does. This is a sign of a large buffer. If the buffer were small as you claim, there would be a packet loss increase but no latency increase.
This is not true. Linksys switched to VxWorks not because of lawsuits (otherwise why continue with the GL?), but because it was perceived to have a lower memory footprint than Linux, with the license cost being less than the cost of the "extra" hardware needed to run Linux. Of course, the DD-WRT crew proved them wrong after a few months by trimming down their distro to fit in the reduced-memory devices while retaining as much functionality as the stock firmware.