#1: You can't just take a barrel of oil and get a barrel of diesel. When oil is processed you get some % as gas and some other % as diesel (and a whole bunch of other things). #2: Gas vs Diesel prices are not just based on the supply of oil. There is additional price variance coming from the demands. A lot of the diesel pricing comes from shipping needs (trucks, trains, etc).
However, Diesel is a good idea because we can more easily convert, transport, and store bio-diesel from plant oils than H2 for fuel cells at this time.
Wow, you are really bad at math(s). I'd much rather have 8 euro change (5+2+1) than have 7.25 euro. (5+2+.20+.05) The.20 euro coin is almost the same size as the 1 euro coin.
The US news outlets love wikileaks. The simple fact is that the news outlets need wikileaks to avoid getting boot-stomped by the US government. "We didn't leak this, we're just reporting it"
I had a similar experience when I wanted to know more about UNIX systems. The funny thing is he continues to be a MS fanboy. This is slightly understandable since he works in business software systems.
I wonder if he still thinks snowboarding is a fad.
Wait, fsfe claims "W3C, the standard setting organization (SSO) that governs the Internet standards". Tthe "World Wide Web Consortium" just works on web stuff. The IETF does Internet standards.
The problem is the Android core has no way to magically know the context for why the app needs access to specific APIs. "Why would this barcode app need access to my contact?" For one you can include contact information in a barcode and it makes it easy to add contacts.
The real problem is that the API access controls are not fine grained enough. The barcode reader app should only have WRITE access to your contacts, not READ.
Users (like me) do NOT want to be continuously prompted for stupid "ARE YOU SURE!?" when running apps on a mobile phone.. See Windows Vista. Agreeing to allow an app data access at install time is by far the least intrusive but still useful method I've seen for data protection I've seen.
This is completely misleading and wrong. The ads system matches content to an ad. At no point is there even a reason for the data to leave Google. If you could sign up for an advertiser account and get data out of Google people would be marching with torches and pitchforks.
There is no conflict between the user serving side and the ads serving side. Especially when you consider that Apps enterprise admins/users can literally turn ads off.
Things at that zoom level are NOT from satellites. They're from aircraft areal photos. These photos only exist where there is an interesting market where someone can sell Google (or bing, or whoever) a license to show them.
Damn, that'd be nice. I pay around 40% of my paycheck between state and federal taxes in the US. I'm sure 10% of my income is paid as health insurance by my employer. The amount of services I get for my nearly similar tax rate is abysmal.
Actual cost for employees is probably higher (after benefits and such), I usually 2x actual salary. That total comes out to about $1M. 150K for a SAN will you get about 10T of usable storage. That comes out to about $100/year/GB. If spread these costs out a bit:
Capitol costs for server equipment: 150+20+80+200 = $450K
Operating costs: Rack space (1 rack @ $1500/month x 36 months) = $54K Vendor service contracts (10% of equipment cost per year) 45 x 3 = $135K Support staff ($200K for 2 senior, 100K for 2 junior) 600 x 3 = $1800K
Grand total cost for 3 years: $2.439M
Total comes to $238/GB or $6.6/GB/Month.
And in reality, a single senior sysadmin/network admin @ 200K/year salary could take care of that amount of equipment and it would take at worst 25% of their work hours. I know because I used to do that kind of thing. So 25% of 200K replacing the 1.8M above would bring the SAN cost down to $77/GB or $2.14/GB/Month.
If you'd be willing to share some basic info, like the domain being crawled, the IP that the traffic came from, and the datestamp, I could have this looked into.
This sorta doesn't work because what you care about in datacenter cooling is maintaining a constant equipment inlet temp. For all practical uses this means your AmbientT and ColdT are the same. What you did get right is that you want the largest delta T in your cooling equipment to provide efficient cooling. No matter what you do with hot or cold "containment" the end goal is to keep the HotT as high as possible when it hits your cooling system.
That's because you have an improperly designed datacenter. A good modern datacenter should have 70-80F inlet temps for most equipment. The problem is if you don't do hot isle containment you have to use super cold air from the chilling equipment to keep that inlet temp requirement near the top of the racks.
I think containing the hot isle is probably the best way to go as well.
* When I'm working in a datacenter I'd rather be walking around in the cold isle (~70-80F in a modern datacenter) than the hot isle (100-120F if properly contained) * Containing the hot isle and to a small space and using the rest of the air and space around the rack (up to the ceiling, walking isles, etc) allows more volume of cool air to be a buffer in case of low/failed cooling capacity.
No, you don't. I helped run the network for a school a while back. We didn't filter anything. We logged everything using a proxy. We simply made it very well known to the students that anything they surfed would be logged. We never had any issues. This was even the school for "bad kids"
We had a couple of the "bad hackers" from the highschool. We made them (with supervision) in charge of keeping the linux machines in the computer lab running.
Yup, my quick launch buttons on my gnome panel are termina, firefox, lock screen. 99% of what I do is either a terminal or a browser window these days. Although I tend to just open them and leave them where they are and use screen or tabs in the browser.
I mostly use ion at work, I just wish it had better gnome applet integration to make it easier to run things like network-manager and that kind of thing.
After thinking about the original question a bit, I think the only actual goal here is to turn off sector re-mapping to prevent absurd seeking when streaming data to/from the disk. But as many other people have posted, most of what is being asked for is completely impossible with modern drives.
No shit, anyone with half a clue has moved to distributed server workloads. I can get much higher processing power out of a stack of 1U machines with 4 drives, not to mention the amount of ram you can get in a cabinet full of those kind of machines. 150T+ disk and 1-2T of ram per rack depending on your machine density.
Who needs a SAN when you can 2x or more replicate your data on a bunch of cheap machines.
I hope so too. Google didn't have federation at first, but eventually it got sorted and now it just works. I'll dance the day Yahoo adds a Jabber federation gateway to their IM.
Slashdot Mirror
There is good and bad with this comment.
#1: You can't just take a barrel of oil and get a barrel of diesel. When oil is processed you get some % as gas and some other % as diesel (and a whole bunch of other things).
#2: Gas vs Diesel prices are not just based on the supply of oil. There is additional price variance coming from the demands. A lot of the diesel pricing comes from shipping needs (trucks, trains, etc).
However, Diesel is a good idea because we can more easily convert, transport, and store bio-diesel from plant oils than H2 for fuel cells at this time.
Wow, you are really bad at math(s). I'd much rather have 8 euro change (5+2+1) than have 7.25 euro. (5+2+.20+.05) The .20 euro coin is almost the same size as the 1 euro coin.
The US news outlets love wikileaks. The simple fact is that the news outlets need wikileaks to avoid getting boot-stomped by the US government. "We didn't leak this, we're just reporting it"
I had a similar experience when I wanted to know more about UNIX systems. The funny thing is he continues to be a MS fanboy. This is slightly understandable since he works in business software systems.
I wonder if he still thinks snowboarding is a fad.
Wait, fsfe claims "W3C, the standard setting organization (SSO) that governs the Internet standards". Tthe "World Wide Web Consortium" just works on web stuff. The IETF does Internet standards.
That's a pretty great example. I'll have to use that to remind the people in my family who claim to be good christians but spout judgmental bullshit.
The problem is the Android core has no way to magically know the context for why the app needs access to specific APIs. "Why would this barcode app need access to my contact?" For one you can include contact information in a barcode and it makes it easy to add contacts.
The real problem is that the API access controls are not fine grained enough. The barcode reader app should only have WRITE access to your contacts, not READ.
Users (like me) do NOT want to be continuously prompted for stupid "ARE YOU SURE!?" when running apps on a mobile phone.. See Windows Vista. Agreeing to allow an app data access at install time is by far the least intrusive but still useful method I've seen for data protection I've seen.
Correlation != Causation.
The only way for an app to get your gmail credentials out of your phone directly is if it asked for your gmail password.
This is completely misleading and wrong. The ads system matches content to an ad. At no point is there even a reason for the data to leave Google. If you could sign up for an advertiser account and get data out of Google people would be marching with torches and pitchforks.
There is no conflict between the user serving side and the ads serving side. Especially when you consider that Apps enterprise admins/users can literally turn ads off.
Things at that zoom level are NOT from satellites. They're from aircraft areal photos. These photos only exist where there is an interesting market where someone can sell Google (or bing, or whoever) a license to show them.
Damn, that'd be nice. I pay around 40% of my paycheck between state and federal taxes in the US. I'm sure 10% of my income is paid as health insurance by my employer. The amount of services I get for my nearly similar tax rate is abysmal.
Of course it's probably not all the costs. It is a good estimate of the basics.
Calling me "ignorant or stupid" without posting any additional/better estimates is just trolling.
Hell, I had 2GB of ram in each of my redundant Fibre controllers over 6 years ago. And that was our TINY 2T system that used 72GB FC drives.
Actual cost for employees is probably higher (after benefits and such), I usually 2x actual salary. That total comes out to about $1M. 150K for a SAN will you get about 10T of usable storage. That comes out to about $100/year/GB. If spread these costs out a bit:
Capitol costs for server equipment:
150+20+80+200 = $450K
Operating costs:
Rack space (1 rack @ $1500/month x 36 months) = $54K
Vendor service contracts (10% of equipment cost per year) 45 x 3 = $135K
Support staff ($200K for 2 senior, 100K for 2 junior) 600 x 3 = $1800K
Grand total cost for 3 years: $2.439M
Total comes to $238/GB or $6.6/GB/Month.
And in reality, a single senior sysadmin/network admin @ 200K/year salary could take care of that amount of equipment and it would take at worst 25% of their work hours. I know because I used to do that kind of thing. So 25% of 200K replacing the 1.8M above would bring the SAN cost down to $77/GB or $2.14/GB/Month.
$30/GB/Month is a scam.
If you'd be willing to share some basic info, like the domain being crawled, the IP that the traffic came from, and the datestamp, I could have this looked into.
This sorta doesn't work because what you care about in datacenter cooling is maintaining a constant equipment inlet temp. For all practical uses this means your AmbientT and ColdT are the same. What you did get right is that you want the largest delta T in your cooling equipment to provide efficient cooling. No matter what you do with hot or cold "containment" the end goal is to keep the HotT as high as possible when it hits your cooling system.
That's because you have an improperly designed datacenter. A good modern datacenter should have 70-80F inlet temps for most equipment. The problem is if you don't do hot isle containment you have to use super cold air from the chilling equipment to keep that inlet temp requirement near the top of the racks.
I think containing the hot isle is probably the best way to go as well.
* When I'm working in a datacenter I'd rather be walking around in the cold isle (~70-80F in a modern datacenter) than the hot isle (100-120F if properly contained)
* Containing the hot isle and to a small space and using the rest of the air and space around the rack (up to the ceiling, walking isles, etc) allows more volume of cool air to be a buffer in case of low/failed cooling capacity.
CIPA didn't exist in 1998.
No, you don't. I helped run the network for a school a while back. We didn't filter anything. We logged everything using a proxy. We simply made it very well known to the students that anything they surfed would be logged. We never had any issues. This was even the school for "bad kids"
We had a couple of the "bad hackers" from the highschool. We made them (with supervision) in charge of keeping the linux machines in the computer lab running.
Yup, my quick launch buttons on my gnome panel are termina, firefox, lock screen. 99% of what I do is either a terminal or a browser window these days. Although I tend to just open them and leave them where they are and use screen or tabs in the browser.
I mostly use ion at work, I just wish it had better gnome applet integration to make it easier to run things like network-manager and that kind of thing.
Try http://www.google.com/adwords
After thinking about the original question a bit, I think the only actual goal here is to turn off sector re-mapping to prevent absurd seeking when streaming data to/from the disk. But as many other people have posted, most of what is being asked for is completely impossible with modern drives.
No shit, anyone with half a clue has moved to distributed server workloads. I can get much higher processing power out of a stack of 1U machines with 4 drives, not to mention the amount of ram you can get in a cabinet full of those kind of machines. 150T+ disk and 1-2T of ram per rack depending on your machine density.
Who needs a SAN when you can 2x or more replicate your data on a bunch of cheap machines.
I hope so too. Google didn't have federation at first, but eventually it got sorted and now it just works. I'll dance the day Yahoo adds a Jabber federation gateway to their IM.