I love Netflix, I have been a continuous member so long that in-fact my rate is still $13.99 and I get 4 DVDs out at once.
I digress, when Netflix first began (or more accurately when I first joined), there was no throttling. When their subscriber base went up they did start "holding back DVDs" for very high rate customers. Basically anyone exceeding 2x the number of DVDs in the plan per month fell into this category. So if you had a 4-out plan, then when you were on your 9th DVD in a 30 day period, the 9th, 10th, and 11th DVDs would take quite sometime to get to you (it would not ship, it would show as available). Now realize I have a national talk radio show so I wouldn't put my name on the line if I wasn't sure. My mother also has a Netflix account and I ordered the SAME DVD one day later (on several test occasions) and it would arrive at her house first.
When this started looking like some really bad PR for Netflix, they fixed it, stocked up on titles, put DCs in cities around the US with the highest Netflix user population, this improved their mailing times, and started researching allowing users to watch movies over the web (which consequently is now deployed). Needless to say, throttling is no more with Netflix, and the customers who do get their monies worth -- Netflix just looses cash on. The rest (the people who let the DVDs sit at home) is how they make their profit.
Some months I too rent and return 30-40 DVDs, others I do 0. And now with no throttling, but yes, it was a reality. Netflix rocks, and Blockbuster is no comparison unless you're willing to give up diversity, a better site, excellent rating and suggestion abilities, for the ability to go grab the latest movie off the shelf without paying $3.99 to walk out of Blockbuster with it (directly) -- not to mention, using YOUR gas to go get the DVD. Think about Blockbuster is 2 miles from your house and you pay $3/gallon then its $0.66 to go grab the DVD from Blockbuster. It's free to get it in your mailbox. Blockbuster is saving themselves money everytime you come in -- and its more expensive than Netflix.
Anyway, just my thoughts and experiences! Take care television and movie lovers.
Moderators and meta-moderators, add the concept of eBay and digg and you've got what you need for Wikipedia. Allow changes to be made immediately. If people agree with what you add, then they say so, by doing so they are given points 'potential points' to moderate with in the future, when a 'trusted' meta-moderator agrees with the moderation the points are actually granted. These points allow someone to gain noteriaty and use their points for saying someone else's additions were incorrect, unfair, etc. Without rating people positively, and then being meta-moderated to create a check and balance, you don't have points to tear people's posts down.
When enough 'trusted' people use their points to 'undo' a modification, it actually becomes undone and is only viewable in a history. If someone has several of their stories taken down by different moderators that are unrelated (by IP/e-mail/etc) then they are prevented from their changes (or changes at their IP) being immediately added to the system and a moderator would view said changes first and approve them.
The community would police itself.:) I might have missed some of my concepts, but I have thought about how Wikipedia could implement this for a long time because there are three types of Wikipedia users:
1) Content getters (read-only types) 2) Content posters (the occasional update/edit/addition) 3) Wikipedia nuts (massive amounts of work and time spent on Wikipedia)
The vast majority is a #1, I happen to be a #2 as many of you likely are, and #3's are just #2's with too much time. This really focuses on helping #1s (and the rest of us) get better information from Wikipedia by the #2s and #3s getting more involved (and encouraging more #1s to become #2/3s)
What's more, Google can only download images and cache images that are A) not prevented from cache in the ROBOTS.TXT, B) accessable to the public, and C) linked from somewhere public it spiders.
There is no magic, it does not magically know where these images are, it does not log in miraculacously, and it does NOT disobey the ROBOTS.TXT file.
Put some protection on the images, move them from the public server to a login only server or use HTTPS (which is not cached), perhaps a ROBOTS.TXT file, and UNLINK the images from the public, even hidden links are visible to Google.
First, let me state that the man was just sharing his opinions for us to meander, at this point what you've have shared with us I would personally consider a bit rude.
Second, this is where I get to say something like you -- if you build computers for a living, get another job. (read on);)
CPUs actually can run at many different temperatures, at many different extremes, remember overclocking? That was playing with death. How hot could you get it before math calculations became random number generators -- that was the old way to do it. (thanks to Josh for that funny tid bit)
AMDs not lasting as long as Intel -- (*stops laughing*) umm do you have the Intel plugged in? I'd believe you, but when was the last time you saw a processor run out of life (without taking off cooling units), and how long did it live? I would venture to say that your motherboard would run out of life before your processor. Last time I checked the periodic table of elements, Silicon half-life was longer than your life. I've still got an XT (pre 286 for those who don't know) that actually runs (linux). But the most important thing here is, in 5 years will your processor still work (assuming you leave the heat sink and fan on till then)? Yes. Do you think that you will still be running a 1.4 GHz machine in 2006? No. So, to conclude this segment, does it really matter? No. (Rhetorical questions are great)
Next, SCSI to Ultra DMA ATA 100 -- decrease in speed, that you are definately correct on. One of my systems is all SCSI and it will rip apart any IDE system drive access time-wise hands down, but the difference between a 5400 and 7200, not as signifigant as you make it out to be. Since you build computers, you know that cache is to a processor like seek time is to a fast hard drive, not platter speed. Platter speed in many cases is directly related to seek time, but not always, especially in IBMs drives. In general IBM's drives are built well, last a long time, and have a good warranty, but even I to bought a 75 GXP and it too failed, TWICE. (Anyone want me to send them a platter, I gave up and made really cool coasters...)
Drive failure of 5400 RPM to 7200 RPM drives is not 1:1 as a matter of fact there was a study published that 5400 RPM drives and 10000 RPM drives did signifigantly better than their 7200 RPM counterparts. There are several factors that directly affect the life of a drive:
-Company Production Quality (e.g. IBM)
-Production / Manufacturing Location
-Drive Line (e.g. 75GXP)*
-BIG ONE: Shipping provider(s) (e.g. UPS, FedEx)
-Spindle speed
-Drive casing temperature rating
THINGS THAT ALSO APPLY THAT PEOPLE DON'T THINK ABOUT:
-How many hard drives are in the same system, how close are they to this hard drive, how much heat are they producing, how much heat is being expunged.
Remember, these are simple uni-directional motors that spin metal platters fast enough to be a deadly weapon or drive your vehicle. Just like a car engine, they can not get too hot before the motor has thermal breakdown (it's an electromagnet).
In the case of the 75GXP from IBM, I know for a fact that this line of drives had major problems, had nothing to do with speed, on the other hand there were many companies who had problems with their 7500 RPM drives, I have a 15000 RPM drive that is hotter than hot, you can cook eggs on that thing and while it specifically states it does not need a fan, they warn you about putting it near other drives.
I agree with you on the PlexWriter or SCSI Yahama, as a matter of fact I have those two drives in my all SCSI system at home -- good picks. For an all IDE system, go with Ricoh -- they have a CD/DVD/CD-R/CD-RW drive, it's nice, not hardware accelerated DVD, but that's what a good graphics card like the ATI Radeon 8100 is for;)
I disagree on Creative sucking tons of CPU power, that is far from true -- they drivers do often suck, but with a bit of skill you don't install any of their software, just the INF file and its supporting drivers and they are great for low noise and low processor usage (assuming you don't install their stupid software) -- Turtle Beach are 'good' cards, but their support is dwendling, and their sound floor is far too high for editing or production work (most also don't support bitrates higher than 44,100 16bit without special software)
Network cards: First, I have a 3905C in my home computer, and I don't know about the specific Intel card that he's talking about, but Intel makes the most advanced NIC cards you can buy. What makes NIC cards better or worse is how much of the OSI model the card supports directly and its interface with memory, the Intel EtherFast PRO 10/100+ that I have at home is a bus-mastering, multi-processor, on-board TCP/IP stack card -- that means to OS doesn't even have to do the protocol work -- now that's pretty cool -- to bad that's just a home computer and not a server with that pimp Intel card... Sheesh.
UPS -- a UPS on ANY system can save a man nightmares, and with the computer market the way it is, they are so cheap there is no excuse not to have at least 30 minutes UPS time on any computer you own... (execept for my XT) Sure you wouldn't be plaing games if the power went out (most likely;) but at least you have time to shut down your computer. I have one computer (my workstation system) that has so much power backup I have no idea how long it will last (it also has two 21' monitors so possibally not long)
But in any case, thank you for sharing with us your secondary opinions.
Time Warner (now part of AOL yuk!), still doing alright in the broadband/cable delivery market, is coming out with a new all-in-one service that is over the truly versitile medium, RG6. The only overcomming that must be made with RG6 (or commonly known as coxial cable) is segregation of RF transmission on the line itself -- once that can be done efficiently (which Time Warner excells at) -- you can get truly amazing speeds (~ +/- 14 Mbps). The question is, do they (or anyone else) have the upstream to support it?
The all-in-one system, which I can't talk too much about due to an impeding NDA, is scheduled for first-quarter next year, but you know how technology is. I am currently testing one that has a beautiful Cisco RG6ethernet/telephone (propritery protocol)/digital cable (encrypted stream)/VOD(video on demand)/VoXip (voice over IP) - routing/bridge (brouter).
They even have plans to have fibre to your house within several years.
Slashdot Mirror
I love Netflix, I have been a continuous member so long that in-fact my rate is still $13.99 and I get 4 DVDs out at once.
I digress, when Netflix first began (or more accurately when I first joined), there was no throttling. When their subscriber base went up they did start "holding back DVDs" for very high rate customers. Basically anyone exceeding 2x the number of DVDs in the plan per month fell into this category. So if you had a 4-out plan, then when you were on your 9th DVD in a 30 day period, the 9th, 10th, and 11th DVDs would take quite sometime to get to you (it would not ship, it would show as available). Now realize I have a national talk radio show so I wouldn't put my name on the line if I wasn't sure. My mother also has a Netflix account and I ordered the SAME DVD one day later (on several test occasions) and it would arrive at her house first.
When this started looking like some really bad PR for Netflix, they fixed it, stocked up on titles, put DCs in cities around the US with the highest Netflix user population, this improved their mailing times, and started researching allowing users to watch movies over the web (which consequently is now deployed). Needless to say, throttling is no more with Netflix, and the customers who do get their monies worth -- Netflix just looses cash on. The rest (the people who let the DVDs sit at home) is how they make their profit.
Some months I too rent and return 30-40 DVDs, others I do 0. And now with no throttling, but yes, it was a reality. Netflix rocks, and Blockbuster is no comparison unless you're willing to give up diversity, a better site, excellent rating and suggestion abilities, for the ability to go grab the latest movie off the shelf without paying $3.99 to walk out of Blockbuster with it (directly) -- not to mention, using YOUR gas to go get the DVD. Think about Blockbuster is 2 miles from your house and you pay $3/gallon then its $0.66 to go grab the DVD from Blockbuster. It's free to get it in your mailbox. Blockbuster is saving themselves money everytime you come in -- and its more expensive than Netflix.
Anyway, just my thoughts and experiences! Take care television and movie lovers.
Moderators and meta-moderators, add the concept of eBay and digg and you've got what you need for Wikipedia. Allow changes to be made immediately. If people agree with what you add, then they say so, by doing so they are given points 'potential points' to moderate with in the future, when a 'trusted' meta-moderator agrees with the moderation the points are actually granted. These points allow someone to gain noteriaty and use their points for saying someone else's additions were incorrect, unfair, etc. Without rating people positively, and then being meta-moderated to create a check and balance, you don't have points to tear people's posts down.
:) I might have missed some of my concepts, but I have thought about how Wikipedia could implement this for a long time because there are three types of Wikipedia users:
When enough 'trusted' people use their points to 'undo' a modification, it actually becomes undone and is only viewable in a history. If someone has several of their stories taken down by different moderators that are unrelated (by IP/e-mail/etc) then they are prevented from their changes (or changes at their IP) being immediately added to the system and a moderator would view said changes first and approve them.
The community would police itself.
1) Content getters (read-only types)
2) Content posters (the occasional update/edit/addition)
3) Wikipedia nuts (massive amounts of work and time spent on Wikipedia)
The vast majority is a #1, I happen to be a #2 as many of you likely are, and #3's are just #2's with too much time. This really focuses on helping #1s (and the rest of us) get better information from Wikipedia by the #2s and #3s getting more involved (and encouraging more #1s to become #2/3s)
This article reminded me I needed to cancel. Done.
What's more, Google can only download images and cache images that are A) not prevented from cache in the ROBOTS.TXT, B) accessable to the public, and C) linked from somewhere public it spiders. There is no magic, it does not magically know where these images are, it does not log in miraculacously, and it does NOT disobey the ROBOTS.TXT file. Put some protection on the images, move them from the public server to a login only server or use HTTPS (which is not cached), perhaps a ROBOTS.TXT file, and UNLINK the images from the public, even hidden links are visible to Google.
I know a guy who has a company that can build these very well priced, check out his website.
The Elegant Universe is a Great book!
Spider-Man is better, you're right!
I've edited my settings (after logging in visit http://subscribe.yahoo.com/showaccount) -- much better :)
First, let me state that the man was just sharing his opinions for us to meander, at this point what you've have shared with us I would personally consider a bit rude.
;)
;)
;) but at least you have time to shut down your computer. I have one computer (my workstation system) that has so much power backup I have no idea how long it will last (it also has two 21' monitors so possibally not long)
Second, this is where I get to say something like you -- if you build computers for a living, get another job. (read on)
CPUs actually can run at many different temperatures, at many different extremes, remember overclocking? That was playing with death. How hot could you get it before math calculations became random number generators -- that was the old way to do it. (thanks to Josh for that funny tid bit)
AMDs not lasting as long as Intel -- (*stops laughing*) umm do you have the Intel plugged in? I'd believe you, but when was the last time you saw a processor run out of life (without taking off cooling units), and how long did it live? I would venture to say that your motherboard would run out of life before your processor. Last time I checked the periodic table of elements, Silicon half-life was longer than your life. I've still got an XT (pre 286 for those who don't know) that actually runs (linux). But the most important thing here is, in 5 years will your processor still work (assuming you leave the heat sink and fan on till then)? Yes. Do you think that you will still be running a 1.4 GHz machine in 2006? No. So, to conclude this segment, does it really matter? No. (Rhetorical questions are great)
Next, SCSI to Ultra DMA ATA 100 -- decrease in speed, that you are definately correct on. One of my systems is all SCSI and it will rip apart any IDE system drive access time-wise hands down, but the difference between a 5400 and 7200, not as signifigant as you make it out to be. Since you build computers, you know that cache is to a processor like seek time is to a fast hard drive, not platter speed. Platter speed in many cases is directly related to seek time, but not always, especially in IBMs drives. In general IBM's drives are built well, last a long time, and have a good warranty, but even I to bought a 75 GXP and it too failed, TWICE. (Anyone want me to send them a platter, I gave up and made really cool coasters...)
Drive failure of 5400 RPM to 7200 RPM drives is not 1:1 as a matter of fact there was a study published that 5400 RPM drives and 10000 RPM drives did signifigantly better than their 7200 RPM counterparts. There are several factors that directly affect the life of a drive:
-Company Production Quality (e.g. IBM)
-Production / Manufacturing Location
-Drive Line (e.g. 75GXP)*
-BIG ONE: Shipping provider(s) (e.g. UPS, FedEx)
-Spindle speed
-Drive casing temperature rating
THINGS THAT ALSO APPLY THAT PEOPLE DON'T THINK ABOUT:
-How many hard drives are in the same system, how close are they to this hard drive, how much heat are they producing, how much heat is being expunged.
Remember, these are simple uni-directional motors that spin metal platters fast enough to be a deadly weapon or drive your vehicle. Just like a car engine, they can not get too hot before the motor has thermal breakdown (it's an electromagnet).
In the case of the 75GXP from IBM, I know for a fact that this line of drives had major problems, had nothing to do with speed, on the other hand there were many companies who had problems with their 7500 RPM drives, I have a 15000 RPM drive that is hotter than hot, you can cook eggs on that thing and while it specifically states it does not need a fan, they warn you about putting it near other drives.
I agree with you on the PlexWriter or SCSI Yahama, as a matter of fact I have those two drives in my all SCSI system at home -- good picks. For an all IDE system, go with Ricoh -- they have a CD/DVD/CD-R/CD-RW drive, it's nice, not hardware accelerated DVD, but that's what a good graphics card like the ATI Radeon 8100 is for
I disagree on Creative sucking tons of CPU power, that is far from true -- they drivers do often suck, but with a bit of skill you don't install any of their software, just the INF file and its supporting drivers and they are great for low noise and low processor usage (assuming you don't install their stupid software) -- Turtle Beach are 'good' cards, but their support is dwendling, and their sound floor is far too high for editing or production work (most also don't support bitrates higher than 44,100 16bit without special software)
Network cards: First, I have a 3905C in my home computer, and I don't know about the specific Intel card that he's talking about, but Intel makes the most advanced NIC cards you can buy. What makes NIC cards better or worse is how much of the OSI model the card supports directly and its interface with memory, the Intel EtherFast PRO 10/100+ that I have at home is a bus-mastering, multi-processor, on-board TCP/IP stack card -- that means to OS doesn't even have to do the protocol work -- now that's pretty cool -- to bad that's just a home computer and not a server with that pimp Intel card... Sheesh.
UPS -- a UPS on ANY system can save a man nightmares, and with the computer market the way it is, they are so cheap there is no excuse not to have at least 30 minutes UPS time on any computer you own... (execept for my XT) Sure you wouldn't be plaing games if the power went out (most likely
But in any case, thank you for sharing with us your secondary opinions.
Time Warner (now part of AOL yuk!), still doing alright in the broadband/cable delivery market, is coming out with a new all-in-one service that is over the truly versitile medium, RG6. The only overcomming that must be made with RG6 (or commonly known as coxial cable) is segregation of RF transmission on the line itself -- once that can be done efficiently (which Time Warner excells at) -- you can get truly amazing speeds (~ +/- 14 Mbps). The question is, do they (or anyone else) have the upstream to support it?
The all-in-one system, which I can't talk too much about due to an impeding NDA, is scheduled for first-quarter next year, but you know how technology is. I am currently testing one that has a beautiful Cisco RG6ethernet/telephone (propritery protocol)/digital cable (encrypted stream)/VOD(video on demand)/VoXip (voice over IP) - routing/bridge (brouter).
They even have plans to have fibre to your house within several years.