Synergy is wonderful. Unfortunately, a lot of corporate IT departments will likely frown upon it for security reasons, which is why I never try to use it when I have both my laptop and desktop turned on at work.
1) Many OSes don't tile windows well. To have two windows properly tiled on a single monitor, you need to minimize everything but those windows and then choose to tile all nonminimized windows. (At least this is my experience with Windows XP) It's faster to just drag one app to your second monitor and maximize it.
2) The aspect ratio of a single 16:9 screen doesn't fit two 4:3 screens well. While for editing Word documents this is not a bad thing (and could be good in fact), for editing PowerPoint documents, images, and Excel spreadsheets, dual 4:3 is better.
3) Moderate sized 4:3 flat panel displays cost a fraction of the price of an Apple 30" display. The Apple 30" display is $1500-2000, 19" 4:3 displays are $200-250 each.
4) Most workers already have their first monitor. Adding a second is cheaper than chucking it and buying a large widescreen, even if that large widescreen were remotely competitive for these purposes with dual 19s in price.
I have a second monitor in my cube, but it's an old beat-up CRT and I don't have the desk space to use it.:(
They pretty much shot themselves in the foot. By imposing severe throttles on BT, they forced BT users to start using extreme measures (such as crypto).
If they had simply imposed some sensible QoS in the first place, dropping BT users to the end of the line but allowing them to use available bandwidth when it was available rather than choking them with an extremely low cap, both ISPs and BT users would be happier. BT users wouldn't have to port-hop and hide, ISPs wouldn't have BT users making everyone lag.
It can, but (usually) does not.:( As a result, it usually becomes an ISP's worst enemy.
That said, throttling it to a fixed cap is bad, but I would not care if an ISP made BT traffic low-priority, as long as they were clear that they did so.
I hate hidden caps (such Cablevision OptimumOffline's "we'll permanently drop your cap down to 150 KB/sec without warning or notification if you use too much upstream" policy.
I've flown AA a number of times in the past few years, always on a plane with power available in some rows, but never in a row that actually had power.
Better to say "Don't virtualize stuff with high I/O unless your virtualization solution paravirtualizes the I/O".
As others have posted, VMWare Server sucks for high-I/O situations, while other solutions (Xen using paravirtualization instead of HW virtualization, and to some degree even Xen with HWV) perform MUCH better in high-I/O situations.
At least a few tips regarding disk I/O: Paravirtualize. Easy with Xen + Linux hosts/guests When paravirtualizing, don't use file-backed disk images (e.g./mnt/virtimages/windowsxp.img). Give each "virtual disk" its own partition. Easy with LVM. For example, my laptop's virtualized Windows partition is/dev/system/windows - under QEMU/KVM it's VERY responsive (with the exception of guest video resolution changes - QEMU+KQEMU handles this well but QEMU+KVM does not. The result is a blank white window with maximum guest OS CPU usage for a few seconds.)
Also, while the above components may be GPL, there are two other issues:
a) Nothing requires you to provide binaries on demand. Still, any time binaries ARE provided, source for those components must be provided, and there HAS been a violation here.
b) Just because the kernel and glibc are GPL doesn't mean that there aren't any closed-source applications. HiSense could comply with the GPL and release source code for all GPL components and anyone wanting to update their system would likely still be SOL because the update is for a closed-source application that runs on the box.
No clue why you got modded -1, it's a good question. Most consumer-grade routers suck.
Look into DD-WRT or a similar "aftermarket firmware" on a compatible router. I suggest the Buffalo WHR-G54S - Cheap ($50 at Circuit City, $43 or so shipped from NewEgg) and fully compatible with DD-WRT.
The problem is not the CPU speed, but the fact that many routers have too small of an ip_conntrack table (or the equivalent if they do not run Linux). DD-WRT lets you bump up the size of that table and decrease the idle connection timeout time. Boom, most common router problem fixed. (No clue why no manufacturer does this... It's not like an extra 512 entries in the table really takes up that much more memory.)
It also lets you prioritize traffice, dumping BitTorrent (or whatever you choose) traffic to the lowest priority. I can run all the BitTorrent I want and never affect any games.:)
99% of Ethernet switches on the market are store-and-forward, and cut-and-forward switches fall back to store-and-forward under heavy load anyway.
In short, you'd spend $1-2k to get a switch that does cut-and-forward instead of store-and-forward, and shave at most 0.6 ms or so off of your latency for 100Base-T. (Less if your game doesn't send packets at the 1500 byte MTU, which I think most games don't, as they don't need to send that much data at a time.)
As far as the FCC is concerned, your speakers are lower on the food chain than a licensed amateur radio operator. Thus you'll never get anywhere with the FCC if you're complaining about a ham operator who is operating legally. (If you read the details of the FCC certification of your speakers, it basically says that if it receives interference from a licensed transmitter, you're SOL, and may not cause harmful interference to licensed users.) Now if you did something to interfere with that ham, there's a good chance the FCC WOULD come down on you.
That said, MOST hams will try to rectify such interference situations if approached about them politely. I know a number of hams who have installed EMI filtering on their neighbors' equipment on their own dime to rectify such situations. Sadly, it's not required for a ham to do this and some won't. A few bad apples can give any group a bad image.:(
Not much use if channels 12, 13, and 14 are allocated in the same way as 1-11 are.
802.11b/g channels (in the 2.4 GHz band) are 11 MHz wide, and spaced 1 MHz apart. This means that most of them overlap each other.
The only nonoverlapping channels are 1, 6, and 11. If allocated in the same manner (1 MHz increments), 12-14 will overlap 11.
The way 802.11 works, two networks using overlapping (but not the same) channels will perform far worse than two networks on the same channel. This is because the collision avoidance algorithm doesn't work as well if not all participants are able to see the headers of potentially conflicting transmissions.
Re:Ruse to sell more motherboards
on
eSATA Connectors
·
· Score: 4, Informative
One correction, one addition: I'm fairly positive that it's 250 MB/sec per lane, not 150 for PCI-E.
In addition, not only is that per-device, it is per-device, per-direction (full duplex, 250MB/sec to the device and 250MB/sec back at the same time)
As to why PCI-E couldn't have been developed back when PCI or AGP were available (rather than incremental steps) - Moore's law. It simply wasn't possible to make silicon capable of handling PCI Express data rates (each lane uses serial communications at 2.5 gbits/sec, which was definately NOT possible with the silicon available back when PCI or AGP were initially developed.)
For those that wonder why PCI-E uses 2.5 Gbit/sec signaling but only transfers 250MB/sec of data, it is because all data is encoded using either 4B5B or 8B10B encoding (I can't remember which of the two), which maps every 4 data bits to 5 signal bits for 4B5B or 8-to-10 for 8B10B. This is done to ensure a minimum number of bit transitions in a given period of time, and also ensure that the signaling has no DC bias. (i.e. equal number of 0s and 1s no matter what the input data is).
I had a friend who was an EE and worked for the USPTO for a while. They were very aggressive in recruiting her, but after six months she was desperate to get out.
As a few nearby posts have said, they are apparently desperate for bodies. They seem to be in a chicken-and-egg situation - while they are understaffed, the reputation for stress and being underpaid makes it hard for them to hire/retain examiners. Inability to hire/retain examiners results in the existing examiners being overstressed.
Sadly, for a LONG time, the USPTO was one of the government's biggest moneymakers but was also one of the most underfunded, as all of their income went to what was basically a "generic" fund allocation pool. I've heard efforts are being made to rectify this (i.e. let the USPTO use the majority of what they bring in rather than sending it elsewhere), which should help make things a bit saner.
Buffalo WHR-G54S. Readily available in brick-and-mortar stores, supported by nearly every distribution that supports WRT54G variants, and cheaper than the mailorder-only GL.
I got mine at Circuit City for $49. Rock solid and works great with DD-WRT.
It's not cheaper than the G-series (unless you compare a GL from Newegg with a G from a more "mainstream" source, you would be better off comparing GL vs. G in price from the same vendor, and in that case, a GL is $62 from Newegg while a G is $49), but it's far harder to obtain.
You're better off getting a Buffalo WHR-G54S. Easier to obtain and cheaper ($49 at Circuit City), 8M RAM/4M ROM like the pre-cost-reduction WRT54G units, and very well supported by DD-WRT.
The Buffalo routers that are supported by DD-WRT have what essentially amounts to an "emergency" TFTP bootloader. In fact, this bootloader is the only way to initially flash DD-WRT onto them.
From what I've read on the forums, the "emergency" TFTP loader is nearly impossible to break. I know I bricked my WHR-G54S once or twice when getting it set up, I just went and performed the initial flash procedure again.
I would personally reccommend the Buffalo WHR-G54S instead. (Not the HP version, that still has some issues with DD-WRT and OpenWRT, or at least DD...)
It's cheaper, easier to obtain (the "L" variants of the Linksys routers are mailorder-only, while Circuit City sells the Buffalos), and as well supported as the Linksys routers by DD-WRT. I'm running DD on mine and love it.
It's also a bit easier to recover a Buffalo WHR-G54S from an accidental "bricking". The emergency TFTP bootloader is nearly impossible to damage.
Also, I recall (but cannot confirm) hearing that there is a coal plant in Utah that releases more radioactive material into the air EACH DAY due to traces of uranium in the coal they burn than TMI has released in its entire lifetime.
Whether or not this is true, I'd rather live 5 miles from a nuclear plant than a coal plant. With a nuclear plant, there is a tiny probability that something might be released into the air that will affect my health. With a coal plant, you KNOW it's happening daily.
Anyone who has read a detailed description of the Chernobyl incident and the events leading up to it would know that it is simply not possible for such an accident to occur in a civilian reactor in the U.S. (For one thing, U.S. civilian reactors don't use a highly flammable substance as their moderator...)
The fact that I would not even trust myself if I were in his situation not to do something "unfair" subconsciously despite putting as much conscious effort into being objective and unbiased means that fundamentally, I cannot trust him. He can try as hard as he can to be objective and unbiased, but the fact remains that someone with as many connections as he does to one entity is going to have an uncontrollable subconscious bias.
In fact, I would say that the "high risk of loss" mantra hurts the tendency of players to team up and band together.
The fact of the matter is that EVE game mechanics breed an environment of distrust and paranoia. Corp thievery is rampant, and as a result corporations rarely trust new members, making it difficult for an intelligent player to reach their full potential. Similarly, it's impossible to group up with a player to "get to know" them - the moment you gang with someone, they can kill you without repercussions, and more often than not, if you gang up with a stranger, they ARE out to kill you and loot your ship's wreck.
That said, if you have an established group of friends that you explicitly trust without any doubts, EVE can be a great game. I play with a bunch of OLD friends that I've known for many years, and love it. But EVE has a lot of potential that is not being realized because CCP is so focused on PvP to the exception of all else.
Slashdot Mirror
This is probably #1 on one of those "You know you've spent too much time on IRC when..." lists.
:(
:(
And I didn't have to slow down either.
|\/|y 133+ i5 +00 57r0|\|g...
Synergy is wonderful. Unfortunately, a lot of corporate IT departments will likely frown upon it for security reasons, which is why I never try to use it when I have both my laptop and desktop turned on at work.
1) Many OSes don't tile windows well. To have two windows properly tiled on a single monitor, you need to minimize everything but those windows and then choose to tile all nonminimized windows. (At least this is my experience with Windows XP) It's faster to just drag one app to your second monitor and maximize it.
:(
2) The aspect ratio of a single 16:9 screen doesn't fit two 4:3 screens well. While for editing Word documents this is not a bad thing (and could be good in fact), for editing PowerPoint documents, images, and Excel spreadsheets, dual 4:3 is better.
3) Moderate sized 4:3 flat panel displays cost a fraction of the price of an Apple 30" display. The Apple 30" display is $1500-2000, 19" 4:3 displays are $200-250 each.
4) Most workers already have their first monitor. Adding a second is cheaper than chucking it and buying a large widescreen, even if that large widescreen were remotely competitive for these purposes with dual 19s in price.
I have a second monitor in my cube, but it's an old beat-up CRT and I don't have the desk space to use it.
They pretty much shot themselves in the foot. By imposing severe throttles on BT, they forced BT users to start using extreme measures (such as crypto).
If they had simply imposed some sensible QoS in the first place, dropping BT users to the end of the line but allowing them to use available bandwidth when it was available rather than choking them with an extremely low cap, both ISPs and BT users would be happier. BT users wouldn't have to port-hop and hide, ISPs wouldn't have BT users making everyone lag.
It can, but (usually) does not. :( As a result, it usually becomes an ISP's worst enemy.
That said, throttling it to a fixed cap is bad, but I would not care if an ISP made BT traffic low-priority, as long as they were clear that they did so.
I hate hidden caps (such Cablevision OptimumOffline's "we'll permanently drop your cap down to 150 KB/sec without warning or notification if you use too much upstream" policy.
I've flown AA a number of times in the past few years, always on a plane with power available in some rows, but never in a row that actually had power.
Are you trying to feed it via DVI?
DVI + limitation of 1080i but not 1080p = lots of compatibility problems.
The previous poster was talking about 1080i component inputs.
It seems to me that there is more evidence supporting his case than yours - maybe your unit is just defective?
UIDs - where smaller is better and yours is just too goddamned large.
I can't remember if it was Office or Windows, but I remember a 40+ floppy stack at some point...
Better to say "Don't virtualize stuff with high I/O unless your virtualization solution paravirtualizes the I/O".
/mnt/virtimages/windowsxp.img). Give each "virtual disk" its own partition. Easy with LVM. For example, my laptop's virtualized Windows partition is /dev/system/windows - under QEMU/KVM it's VERY responsive (with the exception of guest video resolution changes - QEMU+KQEMU handles this well but QEMU+KVM does not. The result is a blank white window with maximum guest OS CPU usage for a few seconds.)
As others have posted, VMWare Server sucks for high-I/O situations, while other solutions (Xen using paravirtualization instead of HW virtualization, and to some degree even Xen with HWV) perform MUCH better in high-I/O situations.
At least a few tips regarding disk I/O:
Paravirtualize. Easy with Xen + Linux hosts/guests
When paravirtualizing, don't use file-backed disk images (e.g.
Promiscuously? Are you implying that there's more than "just friends" going on between Professor_UNIX and Seumas back in the Journal section?
Also, while the above components may be GPL, there are two other issues:
a) Nothing requires you to provide binaries on demand. Still, any time binaries ARE provided, source for those components must be provided, and there HAS been a violation here.
b) Just because the kernel and glibc are GPL doesn't mean that there aren't any closed-source applications. HiSense could comply with the GPL and release source code for all GPL components and anyone wanting to update their system would likely still be SOL because the update is for a closed-source application that runs on the box.
No clue why you got modded -1, it's a good question. Most consumer-grade routers suck.
:)
Look into DD-WRT or a similar "aftermarket firmware" on a compatible router. I suggest the Buffalo WHR-G54S - Cheap ($50 at Circuit City, $43 or so shipped from NewEgg) and fully compatible with DD-WRT.
The problem is not the CPU speed, but the fact that many routers have too small of an ip_conntrack table (or the equivalent if they do not run Linux). DD-WRT lets you bump up the size of that table and decrease the idle connection timeout time. Boom, most common router problem fixed. (No clue why no manufacturer does this... It's not like an extra 512 entries in the table really takes up that much more memory.)
It also lets you prioritize traffice, dumping BitTorrent (or whatever you choose) traffic to the lowest priority. I can run all the BitTorrent I want and never affect any games.
Good luck finding a cut-and-forward switch.
99% of Ethernet switches on the market are store-and-forward, and cut-and-forward switches fall back to store-and-forward under heavy load anyway.
In short, you'd spend $1-2k to get a switch that does cut-and-forward instead of store-and-forward, and shave at most 0.6 ms or so off of your latency for 100Base-T. (Less if your game doesn't send packets at the 1500 byte MTU, which I think most games don't, as they don't need to send that much data at a time.)
Guess what.
:(
As far as the FCC is concerned, your speakers are lower on the food chain than a licensed amateur radio operator. Thus you'll never get anywhere with the FCC if you're complaining about a ham operator who is operating legally. (If you read the details of the FCC certification of your speakers, it basically says that if it receives interference from a licensed transmitter, you're SOL, and may not cause harmful interference to licensed users.) Now if you did something to interfere with that ham, there's a good chance the FCC WOULD come down on you.
That said, MOST hams will try to rectify such interference situations if approached about them politely. I know a number of hams who have installed EMI filtering on their neighbors' equipment on their own dime to rectify such situations. Sadly, it's not required for a ham to do this and some won't. A few bad apples can give any group a bad image.
Not much use if channels 12, 13, and 14 are allocated in the same way as 1-11 are.
802.11b/g channels (in the 2.4 GHz band) are 11 MHz wide, and spaced 1 MHz apart. This means that most of them overlap each other.
The only nonoverlapping channels are 1, 6, and 11. If allocated in the same manner (1 MHz increments), 12-14 will overlap 11.
The way 802.11 works, two networks using overlapping (but not the same) channels will perform far worse than two networks on the same channel. This is because the collision avoidance algorithm doesn't work as well if not all participants are able to see the headers of potentially conflicting transmissions.
One correction, one addition:
I'm fairly positive that it's 250 MB/sec per lane, not 150 for PCI-E.
In addition, not only is that per-device, it is per-device, per-direction (full duplex, 250MB/sec to the device and 250MB/sec back at the same time)
As to why PCI-E couldn't have been developed back when PCI or AGP were available (rather than incremental steps) - Moore's law. It simply wasn't possible to make silicon capable of handling PCI Express data rates (each lane uses serial communications at 2.5 gbits/sec, which was definately NOT possible with the silicon available back when PCI or AGP were initially developed.)
For those that wonder why PCI-E uses 2.5 Gbit/sec signaling but only transfers 250MB/sec of data, it is because all data is encoded using either 4B5B or 8B10B encoding (I can't remember which of the two), which maps every 4 data bits to 5 signal bits for 4B5B or 8-to-10 for 8B10B. This is done to ensure a minimum number of bit transitions in a given period of time, and also ensure that the signaling has no DC bias. (i.e. equal number of 0s and 1s no matter what the input data is).
I had a friend who was an EE and worked for the USPTO for a while. They were very aggressive in recruiting her, but after six months she was desperate to get out.
As a few nearby posts have said, they are apparently desperate for bodies. They seem to be in a chicken-and-egg situation - while they are understaffed, the reputation for stress and being underpaid makes it hard for them to hire/retain examiners. Inability to hire/retain examiners results in the existing examiners being overstressed.
Sadly, for a LONG time, the USPTO was one of the government's biggest moneymakers but was also one of the most underfunded, as all of their income went to what was basically a "generic" fund allocation pool. I've heard efforts are being made to rectify this (i.e. let the USPTO use the majority of what they bring in rather than sending it elsewhere), which should help make things a bit saner.
Buffalo WHR-G54S. Readily available in brick-and-mortar stores, supported by nearly every distribution that supports WRT54G variants, and cheaper than the mailorder-only GL.
I got mine at Circuit City for $49. Rock solid and works great with DD-WRT.
It's not cheaper than the G-series (unless you compare a GL from Newegg with a G from a more "mainstream" source, you would be better off comparing GL vs. G in price from the same vendor, and in that case, a GL is $62 from Newegg while a G is $49), but it's far harder to obtain.
You're better off getting a Buffalo WHR-G54S. Easier to obtain and cheaper ($49 at Circuit City), 8M RAM/4M ROM like the pre-cost-reduction WRT54G units, and very well supported by DD-WRT.
It's also really easy to recover from a bricking.
The Buffalo routers that are supported by DD-WRT have what essentially amounts to an "emergency" TFTP bootloader. In fact, this bootloader is the only way to initially flash DD-WRT onto them.
From what I've read on the forums, the "emergency" TFTP loader is nearly impossible to break. I know I bricked my WHR-G54S once or twice when getting it set up, I just went and performed the initial flash procedure again.
I would personally reccommend the Buffalo WHR-G54S instead. (Not the HP version, that still has some issues with DD-WRT and OpenWRT, or at least DD...)
It's cheaper, easier to obtain (the "L" variants of the Linksys routers are mailorder-only, while Circuit City sells the Buffalos), and as well supported as the Linksys routers by DD-WRT. I'm running DD on mine and love it.
It's also a bit easier to recover a Buffalo WHR-G54S from an accidental "bricking". The emergency TFTP bootloader is nearly impossible to damage.
Also, I recall (but cannot confirm) hearing that there is a coal plant in Utah that releases more radioactive material into the air EACH DAY due to traces of uranium in the coal they burn than TMI has released in its entire lifetime.
Whether or not this is true, I'd rather live 5 miles from a nuclear plant than a coal plant. With a nuclear plant, there is a tiny probability that something might be released into the air that will affect my health. With a coal plant, you KNOW it's happening daily.
Anyone who has read a detailed description of the Chernobyl incident and the events leading up to it would know that it is simply not possible for such an accident to occur in a civilian reactor in the U.S. (For one thing, U.S. civilian reactors don't use a highly flammable substance as their moderator...)
The fact that I would not even trust myself if I were in his situation not to do something "unfair" subconsciously despite putting as much conscious effort into being objective and unbiased means that fundamentally, I cannot trust him. He can try as hard as he can to be objective and unbiased, but the fact remains that someone with as many connections as he does to one entity is going to have an uncontrollable subconscious bias.
In fact, I would say that the "high risk of loss" mantra hurts the tendency of players to team up and band together.
The fact of the matter is that EVE game mechanics breed an environment of distrust and paranoia. Corp thievery is rampant, and as a result corporations rarely trust new members, making it difficult for an intelligent player to reach their full potential. Similarly, it's impossible to group up with a player to "get to know" them - the moment you gang with someone, they can kill you without repercussions, and more often than not, if you gang up with a stranger, they ARE out to kill you and loot your ship's wreck.
That said, if you have an established group of friends that you explicitly trust without any doubts, EVE can be a great game. I play with a bunch of OLD friends that I've known for many years, and love it. But EVE has a lot of potential that is not being realized because CCP is so focused on PvP to the exception of all else.