300GB Seagate SAS 15k rpm HDD for less than 250 dollars a pop
So the 15k Seagate is about $0.83/GB.
Cheap SSD is about $1.25-$1.50 now, good SSD is about $1.50-$2.00 per GB. Sometimes you can get the better SSDs for $1.30/GB now if you catch a price break or buy in bulk.
(But yes, the math sounds funky... and I don't care to figure out where the poster went wrong.)
Sounds about right. It's faster, but there's still some wait time. So about 100x faster, maybe as much as 500x faster. Plus the time to transfer the data since the SSD works in larger block sizes.
So instead of 100-200 IOPS, now you get 5000-10000 IOPS. Maybe 20k on a really good SSD.
To me the cost per GB of an SSD really makes it tough to justify in a lot of cases especially if you want to store large programs like games where you'll need at least a 100GB SSD. However one area place I have started to use low capacity (8 or 16GB) SSDs are in low noise and/or low power environments. If you team them with an ITX Atom board and the right power supply you can build a small computer with no moving parts whatsoever. And the computer will have a very low power usage for applications like HTPCs or network appliances (like firewalls) where the machine might always be powered on.
10k SATA (or SAS) drives do just fine for games or places where you need something for personal use because you're not making money using the system. I have 10k SATAs in my game machine and it's still "fast enough" while leaving the old 7200 RPM drives in the dust.
For work PCs, now that the SSDs are dropping below $2/GB, where you don't need a lot of local storage, the 64/128GB units are cost efficient enough for any user that does heavy file work or multi-tasks. It's not unheard of to pickup good SSDs now for $1.30-$1.40 per GB.
Nearly all of our laptop users now have SSDs, and the desktops are slowly gaining them as we migrate to Win7. For the laptop users, it's a game changer as their 3-4 year old, multi-core unit, suddenly stops feeling like a boat anchor. Sure, new laptops might have 4 cores instead of 2, and those cores might be 30-40% faster, but it was the hard drive slowness that was killing performance for them.
Search for a "high capacity" server, such as the ones from serversdirect. Some of those hold (50) 3.5" drives (2 for the OS, 48 for the data).
144TB SATA setup runs about $26,000. That's with 48x3TB SATA. If you only want 24x3TB SATA, you save about $8100.
Assume RAID-6 across 12 drives (11+spare) at a time and you'd get about 100 TB net space (or you could do it across 16 drives for each and get about 108 TB of net space).
Which is about $0.24/GB.
Figure 2000W of power? So you'd need a hefty circuit and a heavy duty UPS unit.
You can get 512MB SD cards, in bulk, for about $4/card. Also 1GB SD for about $4/card.
Licensing costs, manuf costs and packaging / shipping are fairly significant factor in that. Unlike CD/DVD where you can fit 100 in a tall cake box, the SD cards aren't as easy to handle in bulk.
CD / DVDs aren't at all a reliable media for backups. I wouldn't recommend anyone to do that, especially for financial data. If you need a backup, do it with a USB hard drive for the local one, and also send an off-site backup over the wire. That is, at least 3 copies (if you include the one you are working on).
If you're going to backup to CD/DVD, you have to definitely make at least 2 copies. And you need to include some sort of parity / ECC / recovery data to increase the odds that you can read the files. And stay away from the cheap media (paying a few extra pennies per disk goes a long way). Then you do generational backups and ensure that a particular piece of data is on at least 3 generations before you stop including it in the backups.
I have decade old CD/DVD recordable discs that are still readable. Some of those have sat in a car year after year (in the shade, but still in temperatures that ranged from 0F up to 140F due to being in a confined, sun-baked car with no ventilation). The ones that start to show errors, I just use dd_rescue to make an ISO of what is still readable, pull one of the PAR2 files off of the disk, then use QuickPar or par2 command line to reconstruct the files from the raw data.
(Opticals, being so cheap and easy to use, are great for things like read-only snapshot backups of small content like your filed taxes every year. Toss the paper receipts, paper copy of the filing, plus a CD-R or DVD-R in a paper sleeve into the envelope and shove the whole thing in a drawer. Heck, there's probably enough room on the disk to include that year plus the previous 7 years.)
It's 2011, dammit, why do people still use optical drives?
Because they still have CDs, DVDs laying around that are not available in electronic file format yet? (Until they get ripped.)
DVD/CD writers are about $18-$20. Not a huge expense for something that gives you a lot of flexibility like creating CDs/DVDs or reading one of the zillion CDs/DVDs created in the last 2 decades.
That being said - I don't have an optical in my laptop any longer. I do still have a USB DVD drive that I can attach when needed, but I don't need it on a day to day basis as the desktop has an optical drive. (Not because the laptop doesn't have room for the drive, but because I prefer to have a 2nd internal SATA drive in that location. Thinkpads allow you to do that with a simple item that replaces the optical drive bay with a SATA drive slot.)
$200 for a machine with multiple components that can generally be bought separately is hard price point to break because you start running into the minimum profitable price point for each component.
For instance, go look at 3.5" hard drive prices on NewEgg, sort by price, see whether you can find any below $40 (there is one, the rest are all $40+). That's because for the manuf, the materials that go into making the drive prevent the retail cost from being lower then the $35-$40 price point. Plus there's the money required to stock this thing and move it from the manufacturer to the dealer. And the retailer doesn't want to deal with items that have to be picked and packed where they only make a few pennies on the item. Even in the 2.5" segment, you can't generally find a drive for less then $35-$40.
Laptops suffer the same issue, nobody wants to make & support a unit that sells for much less then about $250-$350.
Cases, $20-$30. Motherboards bottom out at $40. CPUs $40-$45. PSU $10-$15 (cheap cases generally don't include that). DVD drive $20 or so. RAM is about $15. Video cards (unless you find a $40 motherboard with video) are about $30. Plus the OS license (which drives the cost up another $50-$100).
Puts a minimum price at around $250 or so, plus S&H and taxes. But you're left with a really shitty PSU that will probably fry something a year down the road, a case that is a PITA to work with, and a video card from a decade ago.
I prefer not to skimp on the case. Get something good and it will last you through multiple builds, it will be pleasant to work with, it will keep your stuff cool, and it will not be a noisy POS because it will use 80/120/160mm fans. (Which is why I own a lot of Antec Sonata cases.)
PSU is what keeps your PC innards running properly and protected from the nasty stuff on the line voltage (sags, surges, blips, etc. if you don't have a filtering UPS unit). Plus, they sometimes fail spectacularly and the better units do so in a way that protects the innards. Not wise to go bottom of the barrel here, but you don't have super-expensive. Instead of the $40 PSU, go with something in the $60-$70 range that is 80+ efficiency and 450-550W rating.
I always end up around $350-$450 for a basic PC that won't go out of date right away. Minimum of 2GB/4GB RAM, dual-core CPU, good case and PSU, a budget motherboard (but not low-end), etc. And it will have a small amount of expansion possibilities such as 2 DIMM slots open, 1 PCIe 16x slot, maybe 1-2 PCIe 1x/4x slots and room for additional HDs. If you can get away with Linux, you can shave $50 or so off the build cost due to not having an OS license.
Office PCs tend to be more in the $450-$550 range. The higher cost for Windows Professional license, plus a touch more RAM or a tri/quad core CPU drives that up. And the software licenses for office apps, productivity apps, etc. will double that cost at a minimum.
I don't quite get it even with the inscription. Is it a directional explosive?
Yes, it's a directional explosive with lots of little metal balls for shrapnel inside. It has 700 1/8" steel balls, with an effective range of about 50 yds in a 60 degree wide cone that is very flat. Typically command detonated.
See M18 or M18A1 Claymore on various wikis. GlobalSecurity has a chart showing the coverage area.
A good case and power supply is going to run you about $80+$70 (figure $150 total). A good case can easily last you through multiple upgrade cycles and a good PSU keeps everything else inside safe. (A good filtering UPS helps too, but a cheap PSU is just asking for trouble.)
CPU $100, MB $80, RAM $50, DVD $20. That's about $250 for the basics. Plus another $75 for a regular HD which takes us up to $325 for the internals.
A "decent" video card tends to be in the $75 range as a starting point. Cheaper then that and you're bottom of the barrel and going up to $125-$150 is not a bad idea if you truly care about performance.
Then you have to deal with the operating system license, which is $50-$150 depending on what version of Win7 you want and whether you can qualify for the OEM price. The Win7 Pro retail version is $250.
I wouldn't try to build a "gaming" PC for less then about $550-$600. When you start getting lower then that you're really cutting a lot of corners just to hit some magical price point. And going up to about $750 gets you a heck of a lot of power for not much more money (8GB, quad-core CPU, $150 video card).
Now, if you're going absolutely barebones then you can probably get down around $500.
And you can go higher, but I think that puts you out past the "knee" in the bang for the buck graph. (Four years ago, the magic numbers were $900-$1200, so things have come down a bit in price.)
If you're throwing modern PCs out after 4 years, you're a fool.
Anything introduces in the past 4 years is likely to be dual-core and capable of fitting 4GB of RAM. Put a modern OS on that and you have something that will work just fine for regular users for another 4-6 years if no parts die.
We priced out a new Thinkpad T series this week and compared it to what I already have:
- 8GB RAM (my 4-year old T61p has 4GB and can fit 8GB RAM), sure it might be DDR3 instead of DDR2 but that doesn't get you a whole lot of speed.
- 2.5GHz dual-core i3 vs my 2.2GHz Core 2 Duo. A generous estimate is that the i3 is about 25% faster per clock cycle, which means a performance increase of about 40% faster. A quad-core Intel CPU would tip the balance, but not many folks are CPU-constrained any longer once you get dual-core.
- A newer GPU (that's actually worth something - but it's still a GPU crammed into a laptop, so it's not going to be as fast as my year-old GTX 460).
- Win7 vs WinXP, tempting, but I could also just put Win7 on my existing Thinkpad.
- 128GB SSD for +$320, but I already have a 256GB SSD installed.
Computers (barring going to quad/hex cores) are not getting that much faster year after year. The years of 50-80% faster every year are long gone. We're now only seeing 10-20% per-core performance increase every year, with the rest of the improvements coming from increasing the number of cores. But after you get 4-6 cores in your machine, then what? Odds are that you're now struggling to keep even 2 cores busy all the time.
So that 4-year old Thinkpad is still running strong and will probably last another 2-3 years before it gets retired. And if I do the 8GB upgrade and put Win7 on it, it may go another 3-4 years.
Feature requests are ignored. Bugs are ignored. (There are some exceptions, of course, but as a general rule open source projects are disastrous at considering feature requests and bugs that don't have accompanying source code.)
And one of a few things happen:
- Another group forks the project, pays attention to bugs/requests and implements them. Original project withers and dies.
- A competitor arises with a different code base, pays attention to bugs/requests and users flock to the new project.
- Company tries to run roughshod over the community. Community gets up and leaves (and probably forks the source).
- People put up with until one of the above happens.
At least with open source projects there are options. They may not be palatable options to most users but then most of us find distaste in paying Microsoft every year or two for a reskinned Office Suite / Operating System.
Now, I think the NHC kept the "hurricane" classification a bit longer than was justified, but they possibly did that because they KNOW that most people (especially the news media) focus on winds, instead of the REAL danger from a hurricane: flooding. Even if Irene had completely dissipated to little more than a weak tropical depression by the time it hit New Jersey, you'd still have major damage, power outages and loss of life just from the flooding.
There's also the profit angle. Insurance coverage changes (as well as deductibles) depending on whether it's a tropical storm or a hurricane (according to wind speed). So there's money involved here.
The problem will be the storm surge in the subway tunnels flooding them, and a lot of the power conduits. That will take a long time to drain and repair. And Manhattan without power for a week? Not someplace I want to be...
Is Microsoft supposed to keep a database of every sing USB device ever made, just to get an estimate that's slightly more accurate?
No, but you should keep track of every physical device that gets hooked up to the system and track the best transfer rates (files per second and bytes per second). Which would allow you to make better initial estimates. Or track the last 10 biggest (# of files or # of bytes) transfers and average the results. Both of which are very simple algorithms.
If the best files/sec number says 2 hours and the best bytes/sec number says 1 hour, then initially saying that the transfer will take 90 minutes is not that bad of a guess.
Mostly, tracking history is about improving the initial guess. Once you're 30-50% through the transfer, you should have enough history about the current process and you can start doing weighted estimates based on the last 10 seconds, the last 30 seconds, and the last 120 seconds.
If you don't have history for a device, then you don't give an initial estimate until you are 30-60 seconds into the transfer.
Very few people really need a 500GB primary drive (laptops being the exception due to most only having 1 drive slot, but once you use a SSD in a laptop you'll give up on that and keep the bulk files on external drives). It will still be 5-10 years until SSDs are used for bulk storage. For that, prices will have to get below $0.20/GB (give or take 50%). In a desktop, you run with a SSD as the primary and have a few magnetic drives for handling the bulk storage.
There's a large segment of the population who can do just fine with 64GB drives. Those are already flirting with the $100 point, but they still feel a bit limiting for a lot of users. Most users that I deal with will be more comfortable with a 128GB sized SSD, but those are still in the $250-$350 range which is too much for most people. Those who need the 250-300GB drives are also willing to pay current prices for the capacity. (Generally business users where the lack of wasted time due to waiting on the hard drive saves way more then the incidental cost of the SSD over a year.)
It's the same thing we saw with dual-core a few years ago. As long as the CPUs were all $300+, very few people got onto the dual-core or multi-CPU bandwagon. But once dual-core CPUs got below $200, a lot of people no longer saw it as a big ticket item and started switching in droves.
Get those 128GB units below $150 (and ideally below $100) and you're going to see a lot of people switching because the drives are big enough and cheap enough for their needs. They'll gladly be willing to make the sacrifice in capacity in exchange for the responsiveness of the SSD.
Did you not RTFA? Of course you didn't. It's there. Multiple copy operations to the same destination result in a queue-like behavior.
Which is wonderful for magnetic hard drives - and a featured added just in time for the sunset of those magnetic hard drives.
Because probably by the time Win8 ships, even more people will be using SSDs for their primary drives and those don't care about trying to do multiple writes/reads from the same drive at the same time.
The 64GB SSDs are now down around $100. Use that for the boot drive, keep your bulk files on a 2nd, larger, cheaper, magnetic drive.
I still think the magic number is around $1/GB or $0.80/GB for SSD mass adoption. The cheapest drives this year are still in the $1.60/GB range (up to $2/GB for the Intel 320). So we either need one more die shrink or for process improvements to drive the cost down.
The problem with salting: transfer the matter into "security by obscurity".
The only point of per-site (and ideally per-user) salt is so that a pre-computed rainbow table attack does not work. And salt is a semi-public piece of information. It has to be published in enough places through your code that you must assume that the attacker knows how you salt. And often, the salt is stored right along side the password in the password hash.
Without salt: attackers create the huge rainbow table of about a billion passwords and their resulting has values. They then obtain your password hashes. Now they can simply look up the hash value, find it in their massive pre-computed rainbow table and use the corresponding password to login to your site.
Use once - exploit everywhere.
So, you say, let's use a unique salt for our entire site. Now the attacker's precomputed rainbow table is rendered useless. Now they will either create a new rainbow table (which is a lot of work) or start brute forcing against your password hashes (as salt values *have* to be semi-public in order for them to work). The attack takes longer, but if they do find a password, because you used per-site salts, they can instantly scan the list of the other hashes on your site to see whether any other accounts use that exact same password hash.
Now you have (1) broken account which allowed them to instantly crack any other account that happened to use the same password.
Right - now we do per-user salt (which is the proper way).
Rainbow attacks no longer work. And the weakness in one account's password does not automatically expose other accounts that also use the same password. Every account has to be individually brute-forced, which slows the attacker down a lot if you enforce some form of password complexity.
the gross imbalance in Tank/Healer/DPS numbers, leading to 30+ minutes waits to run an instance with a DPS.
Have you ever played a Tank/Healer in a group? For a long period? In a cross-server PUG?
It's a completely thankless and tedious job because people will pull stuff that you don't want pulled, stand in fire (then bitch because you didn't heal them), and generally just act like idiots who need to be told how to wipe their nose. All because the DPS know that they won't see you again, ever, so they can act like assholes and *maybe* get kicked (if the kick timer isn't on cooldown).
That's why there are 30+ minute wait times for DPS. Because the healers and tanks have said, "screw that, I'd rather solo quest or run only with build mates".
(Cross server LFG needs to die. I played as a tank for 2-3 months after Cataclysm launched, ran lots and lots of cross-server LFG until I finally said "screw that" at the attitudes encountered.)
40 man raiding in Vanilla was just rough. It harkened back to the days of the original EverQuest with their 60-80 man raids, spending an hour or two getting everyone there, then spending a few hours downing the content.
Burning Crusade raiding was decent. You had a pair of 10-man raids which were accessible to the smaller guilds and they introduced the Badge of Justice farming in heroics / raids so that you could buy items that just couldn't or wouldn't drop. Unfortunately for the smaller guilds, once you got done with the (2) 10-man raids you were at an absolute wall because there was no way to progress without forming a 25-man raid team. Which, frankly, was like herding cats and wasn't really worth it. The big activities back then were running battlegrounds for PvP gear (you could get pretty much everything from battleground honor, no need to do stupid arena PvP), running heroics (you had to form groups from the same server/faction and you made heavy use of your friend list), doing the weekly raids, or doing the dailies.
Wrath was fairly decent. You finally had a real 10-man progression track as well as the 25-man track, and you had hard-modes which gave you slightly better gear drops. The big downside in the raid scene was that itemization SUCKED HARD. There was just a real lack of options for a lot of slots in Wrath. Plus Naxx was too much of a pushover, even for a starting raid. Then you had the disaster that was ToGC which was basically "everyone stands in a box and they throw monsters at you" - which was very boring and had none of the charm of a dungeon like Karazhan. Wrath also introduced the concept of cross-server LFG, which destroyed any need to build a friend list to find people to run with (in fact, they broke the same-server LFG tool completely). PvP became elitist where the only way to get current gear was to compete (and WIN) in the arena. Latecomers to the PvP arena party were at a severe disadvantage in gear and could never manage to win, doing not much more then buffing the scores of the early adopters who already had their gear.
So, even going into the tail end of Wrath - a lot of us gave up after ToGC hit because it was bland and boring and uninspired. The final raid of the expansion (Icecrown) was at least somewhat interesting, even if some of the fights were stupid-difficult or stupid-easy.
The other big problem with Wrath was the focus in the upper-end zones of "zoning" where once you completed a quest chain, you were in a separate instance of the area. You could no longer see or interact with anyone who had not completed the quest for that area. Nor could you go back and help fellow guild members, even if grouped. So unless you did the quest chains in lockstep fashion, you became distanced from your quest partner if you got out of sync.
Now, take everything of the above - add it to Cataclysm. Which was a good expansion in that it revamped the old world but you were still stuck with the asinine cross-server LFG (people behaving mean towards others, on purpose, because you'll never see them again), PvP that focused on arena play being the only way to get the top-tier gear, excessive use of the dynamic zoning and solitary quest chains. And you ended up with a situation where a lot of us just didn't find it fun any more. (Cross-server LFG really destroyed the sense of community on the realms where you *had* to network in order to get groups consistently.)
That and the sparkly ponies and critters that could only be bought with real world cash (not even a rare loot drop from a dungeon goodie bag or something).
Weapon skills, a weapon near prefect condition and aiming at the weak points.
Regular super mutants die easily, the brutes are a bit harder but a well maintained Chinese Assault Rifle still slices them up.
Overlords and behemoths require tenderizing first with the use of frag mines, grenades, bottlecap mines.
(Things were actually better balanced in New Vegas - the sad thing is that Obsidian doesn't grasp the concept of "open world" so New Vegas was extremely linear.)
VATS issues were more one of "sometimes it glitches and takes forever to play the animation" or "sometimes it glitches and buries the body in the terrain".
I used it a lot at first, but over time started using it less and less and just started aiming for the center of mass and squeezing off a burst of automatic fire. Or dug out the trusty sniper rifle and did it the easy way.
(One thing that Fallout New Vegas got right was the notch in weapon condition where your weapon would perform exactly as expected above that level of condition. In FO:3, you constantly had to carry spare sniper rifles and spare assault rifles to deal with a weapon that became incredibly ineffective below 75% condition.)
Slashdot Mirror
300GB Seagate SAS 15k rpm HDD for less than 250 dollars a pop
So the 15k Seagate is about $0.83/GB.
Cheap SSD is about $1.25-$1.50 now, good SSD is about $1.50-$2.00 per GB. Sometimes you can get the better SSDs for $1.30/GB now if you catch a price break or buy in bulk.
(But yes, the math sounds funky... and I don't care to figure out where the poster went wrong.)
Sounds about right. It's faster, but there's still some wait time. So about 100x faster, maybe as much as 500x faster. Plus the time to transfer the data since the SSD works in larger block sizes.
So instead of 100-200 IOPS, now you get 5000-10000 IOPS. Maybe 20k on a really good SSD.
To me the cost per GB of an SSD really makes it tough to justify in a lot of cases especially if you want to store large programs like games where you'll need at least a 100GB SSD. However one area place I have started to use low capacity (8 or 16GB) SSDs are in low noise and/or low power environments. If you team them with an ITX Atom board and the right power supply you can build a small computer with no moving parts whatsoever. And the computer will have a very low power usage for applications like HTPCs or network appliances (like firewalls) where the machine might always be powered on.
10k SATA (or SAS) drives do just fine for games or places where you need something for personal use because you're not making money using the system. I have 10k SATAs in my game machine and it's still "fast enough" while leaving the old 7200 RPM drives in the dust.
For work PCs, now that the SSDs are dropping below $2/GB, where you don't need a lot of local storage, the 64/128GB units are cost efficient enough for any user that does heavy file work or multi-tasks. It's not unheard of to pickup good SSDs now for $1.30-$1.40 per GB.
Nearly all of our laptop users now have SSDs, and the desktops are slowly gaining them as we migrate to Win7. For the laptop users, it's a game changer as their 3-4 year old, multi-core unit, suddenly stops feeling like a boat anchor. Sure, new laptops might have 4 cores instead of 2, and those cores might be 30-40% faster, but it was the hard drive slowness that was killing performance for them.
Search for a "high capacity" server, such as the ones from serversdirect. Some of those hold (50) 3.5" drives (2 for the OS, 48 for the data).
144TB SATA setup runs about $26,000. That's with 48x3TB SATA. If you only want 24x3TB SATA, you save about $8100.
Assume RAID-6 across 12 drives (11+spare) at a time and you'd get about 100 TB net space (or you could do it across 16 drives for each and get about 108 TB of net space).
Which is about $0.24/GB.
Figure 2000W of power? So you'd need a hefty circuit and a heavy duty UPS unit.
You can get 512MB SD cards, in bulk, for about $4/card. Also 1GB SD for about $4/card.
Licensing costs, manuf costs and packaging / shipping are fairly significant factor in that. Unlike CD/DVD where you can fit 100 in a tall cake box, the SD cards aren't as easy to handle in bulk.
CD / DVDs aren't at all a reliable media for backups. I wouldn't recommend anyone to do that, especially for financial data. If you need a backup, do it with a USB hard drive for the local one, and also send an off-site backup over the wire. That is, at least 3 copies (if you include the one you are working on).
If you're going to backup to CD/DVD, you have to definitely make at least 2 copies. And you need to include some sort of parity / ECC / recovery data to increase the odds that you can read the files. And stay away from the cheap media (paying a few extra pennies per disk goes a long way). Then you do generational backups and ensure that a particular piece of data is on at least 3 generations before you stop including it in the backups.
I have decade old CD/DVD recordable discs that are still readable. Some of those have sat in a car year after year (in the shade, but still in temperatures that ranged from 0F up to 140F due to being in a confined, sun-baked car with no ventilation). The ones that start to show errors, I just use dd_rescue to make an ISO of what is still readable, pull one of the PAR2 files off of the disk, then use QuickPar or par2 command line to reconstruct the files from the raw data.
(Opticals, being so cheap and easy to use, are great for things like read-only snapshot backups of small content like your filed taxes every year. Toss the paper receipts, paper copy of the filing, plus a CD-R or DVD-R in a paper sleeve into the envelope and shove the whole thing in a drawer. Heck, there's probably enough room on the disk to include that year plus the previous 7 years.)
It's 2011, dammit, why do people still use optical drives?
Because they still have CDs, DVDs laying around that are not available in electronic file format yet? (Until they get ripped.)
DVD/CD writers are about $18-$20. Not a huge expense for something that gives you a lot of flexibility like creating CDs/DVDs or reading one of the zillion CDs/DVDs created in the last 2 decades.
That being said - I don't have an optical in my laptop any longer. I do still have a USB DVD drive that I can attach when needed, but I don't need it on a day to day basis as the desktop has an optical drive. (Not because the laptop doesn't have room for the drive, but because I prefer to have a 2nd internal SATA drive in that location. Thinkpads allow you to do that with a simple item that replaces the optical drive bay with a SATA drive slot.)
$200 for a machine with multiple components that can generally be bought separately is hard price point to break because you start running into the minimum profitable price point for each component.
For instance, go look at 3.5" hard drive prices on NewEgg, sort by price, see whether you can find any below $40 (there is one, the rest are all $40+). That's because for the manuf, the materials that go into making the drive prevent the retail cost from being lower then the $35-$40 price point. Plus there's the money required to stock this thing and move it from the manufacturer to the dealer. And the retailer doesn't want to deal with items that have to be picked and packed where they only make a few pennies on the item. Even in the 2.5" segment, you can't generally find a drive for less then $35-$40.
Laptops suffer the same issue, nobody wants to make & support a unit that sells for much less then about $250-$350.
Cases, $20-$30. Motherboards bottom out at $40. CPUs $40-$45. PSU $10-$15 (cheap cases generally don't include that). DVD drive $20 or so. RAM is about $15. Video cards (unless you find a $40 motherboard with video) are about $30. Plus the OS license (which drives the cost up another $50-$100).
Puts a minimum price at around $250 or so, plus S&H and taxes. But you're left with a really shitty PSU that will probably fry something a year down the road, a case that is a PITA to work with, and a video card from a decade ago.
I prefer not to skimp on the case. Get something good and it will last you through multiple builds, it will be pleasant to work with, it will keep your stuff cool, and it will not be a noisy POS because it will use 80/120/160mm fans. (Which is why I own a lot of Antec Sonata cases.)
PSU is what keeps your PC innards running properly and protected from the nasty stuff on the line voltage (sags, surges, blips, etc. if you don't have a filtering UPS unit). Plus, they sometimes fail spectacularly and the better units do so in a way that protects the innards. Not wise to go bottom of the barrel here, but you don't have super-expensive. Instead of the $40 PSU, go with something in the $60-$70 range that is 80+ efficiency and 450-550W rating.
I always end up around $350-$450 for a basic PC that won't go out of date right away. Minimum of 2GB/4GB RAM, dual-core CPU, good case and PSU, a budget motherboard (but not low-end), etc. And it will have a small amount of expansion possibilities such as 2 DIMM slots open, 1 PCIe 16x slot, maybe 1-2 PCIe 1x/4x slots and room for additional HDs. If you can get away with Linux, you can shave $50 or so off the build cost due to not having an OS license.
Office PCs tend to be more in the $450-$550 range. The higher cost for Windows Professional license, plus a touch more RAM or a tri/quad core CPU drives that up. And the software licenses for office apps, productivity apps, etc. will double that cost at a minimum.
I don't quite get it even with the inscription. Is it a directional explosive?
Yes, it's a directional explosive with lots of little metal balls for shrapnel inside. It has 700 1/8" steel balls, with an effective range of about 50 yds in a 60 degree wide cone that is very flat. Typically command detonated.
See M18 or M18A1 Claymore on various wikis. GlobalSecurity has a chart showing the coverage area.
$450 would be pushing it for a gaming rig.
A good case and power supply is going to run you about $80+$70 (figure $150 total). A good case can easily last you through multiple upgrade cycles and a good PSU keeps everything else inside safe. (A good filtering UPS helps too, but a cheap PSU is just asking for trouble.)
CPU $100, MB $80, RAM $50, DVD $20. That's about $250 for the basics. Plus another $75 for a regular HD which takes us up to $325 for the internals.
A "decent" video card tends to be in the $75 range as a starting point. Cheaper then that and you're bottom of the barrel and going up to $125-$150 is not a bad idea if you truly care about performance.
Then you have to deal with the operating system license, which is $50-$150 depending on what version of Win7 you want and whether you can qualify for the OEM price. The Win7 Pro retail version is $250.
I wouldn't try to build a "gaming" PC for less then about $550-$600. When you start getting lower then that you're really cutting a lot of corners just to hit some magical price point. And going up to about $750 gets you a heck of a lot of power for not much more money (8GB, quad-core CPU, $150 video card).
Now, if you're going absolutely barebones then you can probably get down around $500.
And you can go higher, but I think that puts you out past the "knee" in the bang for the buck graph. (Four years ago, the magic numbers were $900-$1200, so things have come down a bit in price.)
If you're throwing modern PCs out after 4 years, you're a fool.
Anything introduces in the past 4 years is likely to be dual-core and capable of fitting 4GB of RAM. Put a modern OS on that and you have something that will work just fine for regular users for another 4-6 years if no parts die.
We priced out a new Thinkpad T series this week and compared it to what I already have:
- 8GB RAM (my 4-year old T61p has 4GB and can fit 8GB RAM), sure it might be DDR3 instead of DDR2 but that doesn't get you a whole lot of speed.
- 2.5GHz dual-core i3 vs my 2.2GHz Core 2 Duo. A generous estimate is that the i3 is about 25% faster per clock cycle, which means a performance increase of about 40% faster. A quad-core Intel CPU would tip the balance, but not many folks are CPU-constrained any longer once you get dual-core.
- A newer GPU (that's actually worth something - but it's still a GPU crammed into a laptop, so it's not going to be as fast as my year-old GTX 460).
- Win7 vs WinXP, tempting, but I could also just put Win7 on my existing Thinkpad.
- 128GB SSD for +$320, but I already have a 256GB SSD installed.
Computers (barring going to quad/hex cores) are not getting that much faster year after year. The years of 50-80% faster every year are long gone. We're now only seeing 10-20% per-core performance increase every year, with the rest of the improvements coming from increasing the number of cores. But after you get 4-6 cores in your machine, then what? Odds are that you're now struggling to keep even 2 cores busy all the time.
So that 4-year old Thinkpad is still running strong and will probably last another 2-3 years before it gets retired. And if I do the 8GB upgrade and put Win7 on it, it may go another 3-4 years.
Feature requests are ignored. Bugs are ignored. (There are some exceptions, of course, but as a general rule open source projects are disastrous at considering feature requests and bugs that don't have accompanying source code.)
And one of a few things happen:
- Another group forks the project, pays attention to bugs/requests and implements them. Original project withers and dies.
- A competitor arises with a different code base, pays attention to bugs/requests and users flock to the new project.
- Company tries to run roughshod over the community. Community gets up and leaves (and probably forks the source).
- People put up with until one of the above happens.
At least with open source projects there are options. They may not be palatable options to most users but then most of us find distaste in paying Microsoft every year or two for a reskinned Office Suite / Operating System.
Try "nihilist". Or just "self-centered bastard".
Now, I think the NHC kept the "hurricane" classification a bit longer than was justified, but they possibly did that because they KNOW that most people (especially the news media) focus on winds, instead of the REAL danger from a hurricane: flooding. Even if Irene had completely dissipated to little more than a weak tropical depression by the time it hit New Jersey, you'd still have major damage, power outages and loss of life just from the flooding.
There's also the profit angle. Insurance coverage changes (as well as deductibles) depending on whether it's a tropical storm or a hurricane (according to wind speed). So there's money involved here.
The problem will be the storm surge in the subway tunnels flooding them, and a lot of the power conduits. That will take a long time to drain and repair. And Manhattan without power for a week? Not someplace I want to be...
Probably equivalent to the 2003 Northeast US blackout, which lasted for 36-48 hours in most areas.
That was completely unexpected, this will have allowed ConEd and LIPA to plan for it to at least some degree.
(Planning here for 24-72h without power.)
I get a solid 50-55 MB/s transferring files to/from a Samba file share in Win7 64bit.
And the bottleneck there is the hard drive on the server (which is maxed out at 100% utilization).
Naturally, unless I have multiple spindles in a RAID-10 array, I'm not going to try pushing 2 large transfers at once to that set of disks.
Is Microsoft supposed to keep a database of every sing USB device ever made, just to get an estimate that's slightly more accurate?
No, but you should keep track of every physical device that gets hooked up to the system and track the best transfer rates (files per second and bytes per second). Which would allow you to make better initial estimates. Or track the last 10 biggest (# of files or # of bytes) transfers and average the results. Both of which are very simple algorithms.
If the best files/sec number says 2 hours and the best bytes/sec number says 1 hour, then initially saying that the transfer will take 90 minutes is not that bad of a guess.
Mostly, tracking history is about improving the initial guess. Once you're 30-50% through the transfer, you should have enough history about the current process and you can start doing weighted estimates based on the last 10 seconds, the last 30 seconds, and the last 120 seconds.
If you don't have history for a device, then you don't give an initial estimate until you are 30-60 seconds into the transfer.
Very few people really need a 500GB primary drive (laptops being the exception due to most only having 1 drive slot, but once you use a SSD in a laptop you'll give up on that and keep the bulk files on external drives). It will still be 5-10 years until SSDs are used for bulk storage. For that, prices will have to get below $0.20/GB (give or take 50%). In a desktop, you run with a SSD as the primary and have a few magnetic drives for handling the bulk storage.
There's a large segment of the population who can do just fine with 64GB drives. Those are already flirting with the $100 point, but they still feel a bit limiting for a lot of users. Most users that I deal with will be more comfortable with a 128GB sized SSD, but those are still in the $250-$350 range which is too much for most people. Those who need the 250-300GB drives are also willing to pay current prices for the capacity. (Generally business users where the lack of wasted time due to waiting on the hard drive saves way more then the incidental cost of the SSD over a year.)
It's the same thing we saw with dual-core a few years ago. As long as the CPUs were all $300+, very few people got onto the dual-core or multi-CPU bandwagon. But once dual-core CPUs got below $200, a lot of people no longer saw it as a big ticket item and started switching in droves.
Get those 128GB units below $150 (and ideally below $100) and you're going to see a lot of people switching because the drives are big enough and cheap enough for their needs. They'll gladly be willing to make the sacrifice in capacity in exchange for the responsiveness of the SSD.
Did you not RTFA? Of course you didn't. It's there. Multiple copy operations to the same destination result in a queue-like behavior.
Which is wonderful for magnetic hard drives - and a featured added just in time for the sunset of those magnetic hard drives.
Because probably by the time Win8 ships, even more people will be using SSDs for their primary drives and those don't care about trying to do multiple writes/reads from the same drive at the same time.
The 64GB SSDs are now down around $100. Use that for the boot drive, keep your bulk files on a 2nd, larger, cheaper, magnetic drive.
I still think the magic number is around $1/GB or $0.80/GB for SSD mass adoption. The cheapest drives this year are still in the $1.60/GB range (up to $2/GB for the Intel 320). So we either need one more die shrink or for process improvements to drive the cost down.
The problem with salting: transfer the matter into "security by obscurity".
The only point of per-site (and ideally per-user) salt is so that a pre-computed rainbow table attack does not work. And salt is a semi-public piece of information. It has to be published in enough places through your code that you must assume that the attacker knows how you salt. And often, the salt is stored right along side the password in the password hash.
Without salt: attackers create the huge rainbow table of about a billion passwords and their resulting has values. They then obtain your password hashes. Now they can simply look up the hash value, find it in their massive pre-computed rainbow table and use the corresponding password to login to your site.
Use once - exploit everywhere.
So, you say, let's use a unique salt for our entire site. Now the attacker's precomputed rainbow table is rendered useless. Now they will either create a new rainbow table (which is a lot of work) or start brute forcing against your password hashes (as salt values *have* to be semi-public in order for them to work). The attack takes longer, but if they do find a password, because you used per-site salts, they can instantly scan the list of the other hashes on your site to see whether any other accounts use that exact same password hash.
Now you have (1) broken account which allowed them to instantly crack any other account that happened to use the same password.
Right - now we do per-user salt (which is the proper way). Rainbow attacks no longer work. And the weakness in one account's password does not automatically expose other accounts that also use the same password. Every account has to be individually brute-forced, which slows the attacker down a lot if you enforce some form of password complexity.
the gross imbalance in Tank/Healer/DPS numbers, leading to 30+ minutes waits to run an instance with a DPS.
Have you ever played a Tank/Healer in a group? For a long period? In a cross-server PUG?
It's a completely thankless and tedious job because people will pull stuff that you don't want pulled, stand in fire (then bitch because you didn't heal them), and generally just act like idiots who need to be told how to wipe their nose. All because the DPS know that they won't see you again, ever, so they can act like assholes and *maybe* get kicked (if the kick timer isn't on cooldown).
That's why there are 30+ minute wait times for DPS. Because the healers and tanks have said, "screw that, I'd rather solo quest or run only with build mates".
(Cross server LFG needs to die. I played as a tank for 2-3 months after Cataclysm launched, ran lots and lots of cross-server LFG until I finally said "screw that" at the attitudes encountered.)
Your take is definitely a bit off.
40 man raiding in Vanilla was just rough. It harkened back to the days of the original EverQuest with their 60-80 man raids, spending an hour or two getting everyone there, then spending a few hours downing the content.
Burning Crusade raiding was decent. You had a pair of 10-man raids which were accessible to the smaller guilds and they introduced the Badge of Justice farming in heroics / raids so that you could buy items that just couldn't or wouldn't drop. Unfortunately for the smaller guilds, once you got done with the (2) 10-man raids you were at an absolute wall because there was no way to progress without forming a 25-man raid team. Which, frankly, was like herding cats and wasn't really worth it. The big activities back then were running battlegrounds for PvP gear (you could get pretty much everything from battleground honor, no need to do stupid arena PvP), running heroics (you had to form groups from the same server/faction and you made heavy use of your friend list), doing the weekly raids, or doing the dailies.
Wrath was fairly decent. You finally had a real 10-man progression track as well as the 25-man track, and you had hard-modes which gave you slightly better gear drops. The big downside in the raid scene was that itemization SUCKED HARD. There was just a real lack of options for a lot of slots in Wrath. Plus Naxx was too much of a pushover, even for a starting raid. Then you had the disaster that was ToGC which was basically "everyone stands in a box and they throw monsters at you" - which was very boring and had none of the charm of a dungeon like Karazhan. Wrath also introduced the concept of cross-server LFG, which destroyed any need to build a friend list to find people to run with (in fact, they broke the same-server LFG tool completely). PvP became elitist where the only way to get current gear was to compete (and WIN) in the arena. Latecomers to the PvP arena party were at a severe disadvantage in gear and could never manage to win, doing not much more then buffing the scores of the early adopters who already had their gear.
So, even going into the tail end of Wrath - a lot of us gave up after ToGC hit because it was bland and boring and uninspired. The final raid of the expansion (Icecrown) was at least somewhat interesting, even if some of the fights were stupid-difficult or stupid-easy.
The other big problem with Wrath was the focus in the upper-end zones of "zoning" where once you completed a quest chain, you were in a separate instance of the area. You could no longer see or interact with anyone who had not completed the quest for that area. Nor could you go back and help fellow guild members, even if grouped. So unless you did the quest chains in lockstep fashion, you became distanced from your quest partner if you got out of sync.
Now, take everything of the above - add it to Cataclysm. Which was a good expansion in that it revamped the old world but you were still stuck with the asinine cross-server LFG (people behaving mean towards others, on purpose, because you'll never see them again), PvP that focused on arena play being the only way to get the top-tier gear, excessive use of the dynamic zoning and solitary quest chains. And you ended up with a situation where a lot of us just didn't find it fun any more. (Cross-server LFG really destroyed the sense of community on the realms where you *had* to network in order to get groups consistently.)
That and the sparkly ponies and critters that could only be bought with real world cash (not even a rare loot drop from a dungeon goodie bag or something).
Weapon skills, a weapon near prefect condition and aiming at the weak points.
Regular super mutants die easily, the brutes are a bit harder but a well maintained Chinese Assault Rifle still slices them up.
Overlords and behemoths require tenderizing first with the use of frag mines, grenades, bottlecap mines.
(Things were actually better balanced in New Vegas - the sad thing is that Obsidian doesn't grasp the concept of "open world" so New Vegas was extremely linear.)
VATS issues were more one of "sometimes it glitches and takes forever to play the animation" or "sometimes it glitches and buries the body in the terrain".
I used it a lot at first, but over time started using it less and less and just started aiming for the center of mass and squeezing off a burst of automatic fire. Or dug out the trusty sniper rifle and did it the easy way.
(One thing that Fallout New Vegas got right was the notch in weapon condition where your weapon would perform exactly as expected above that level of condition. In FO:3, you constantly had to carry spare sniper rifles and spare assault rifles to deal with a weapon that became incredibly ineffective below 75% condition.)
There are far too many assholes that use spamcop to flex their e-peen and put others through the horror of getting off of various block lists.