If you're a Java addict that's forced to write C++, then you'll do the stupid thing and allocate all your objects with new foo(). Needless to say, this is painfully stupid. If you're smart, you'll normally use RAII and not worry about cleaning up your objects, because they're on the stack.
PCI is nowhere close to being fast enough for USB 3, USB 2 sure, but not USB 3. Also, even a single 7200 RPM SATA hard drive can outstrip the bandwidth provided via a PCI slot nowadays. On the other hand, PCIe is a totally different story, and just about every motherboard these days includes at least a couple PCIe slots.
I think the GP was talking about Intel pricing, where you're pretty much guaranteed to get fleeced. AMD is great on performance/$, but they don't have any chips that compete above the mid to mid-high performance market.
As for IE, some people would just rather not have to deal with third party crap. However, IE6? There really is no reason to run that anymore. Won't IE7 or 8 run on Windows 98?
No, IE7 and IE8 aren't supported on 98, ME or even 2000, XP and newer only.
You must be doing something wrong... Even on a huge story (like this one, 801 comments when I loaded it), Konqueror loaded and rendered all of the comments in a second or two at most. Admittedly, this machine has an i7 in it, but I can't believe that the chip in the N900 is two orders of magnitude slower.
Everybody is trying to improve their Javascript execution speed, so it's really not a big slip. Really, you can't blame them, after all, it lets you play Javascript games like Game! that much faster!
I don't. I watch HDTV on a 3 gigahertz single core Pentium
>>>What do you need 12G of RAM for?
I don't. I only have half-a-gig.
>>>What do you need a 3T hard drive for?
I don't. It's only 0.3 terabytes.
The questions aren't directed at anyone in particular, they're just today's equivalent of the same questions people could have asked 10 years ago... What do you need a 3 GHz Pentium for? What do you need 512M RAM for? What do you need a 300G hard drive for? Of course, the answers to those questions are blindingly obvious now, but they weren't then.
What I'm really trying to point out here is the absurdity of the typical kneejerk reaction of "Oh, there's nothing yet that requires $newtech, therefore it's stupid and nobody should buy it."
Finally, if you're going to go by the literal definition of "need", then your original question is loaded, as there's no correct answer. Nobody needs an Internet connection, a computer, anything electronic or even a home.
What do you need a 3 GHz 6 core CPU for? What do you need 12G of RAM for? What do you need a 3T hard drive for? These are all equally pointless questions, because regardless of the fact that you can't think of anything that would use the faster hardware, there's always countless ideas that would become practical (and widely implemented) when faster hardware is deployed.
And as for value, a good 128GB SSD is $300. For about $200 more, you can get 3 x 150GB Raptors and a $100 Adaptec SATA RAID controller, config it in RAID 5 and get comparable performance, not to mention a little redundancy. The extra initial investement will pay for itself in uptime over the long-term.
I'm sorry, but you're completely and hopelessly wrong. Spinning rust gets around 100 IOPS, maybe 200 at 15k RPM. The Intel X25-E gets around 10,000 IOPS. Assuming linear speedup (which you won't get anything close to), you'd need 100 rotational drives to come close to the performance of a single X25-E.
The only performance metric where SSDs and spinning rust are anywhere close is on linear read/write speeds. Sadly, that's of no consequence, because that workload only exists in benchmarks.
(Also, god help you if you put a database server on RAID 5... goodbye performance! RAID 10 or bust.)
SSDs absolutely blow spinning rust out of the water on a price/performance basis, even the initial models that cost a grand for <100G of storage. The only problem is that there's a lot of poor quality SSDs out there now that perform badly on random writes. If you balk at the price, you don't need the performance anyways, move along.
I think it's revealing that neither Apple nor Google chose to invest in Mozilla instead of going it on their own...
Except they didn't go it on their own. Apple forked KHTML into Webkit because KHTML was (and is) a substantially cleaner codebase than Gecko was (and is). Similarly, Google just forked Webkit instead of Gecko.
That's pretty much the best advise you can get — never trust the client. But as with everything in technology, there's no silver bullet. You also need to consider what you're passing back from the client, try to pass as little as possible, the less you pass back, the less you need to validate and the less chances you have to make mistakes. Also consider what form the data you're passing the data back from the client is in... validating integers is trivial, validating freeform text (especially if you're going to be displaying it somewhere on a page!) isn't.
In the case where you are displaying freeform text from the user (and you'll be hard pressed to avoid this), there are a lot of security problems you need to consider. Ultimately, if you don't let the user use HTML (ever), you can avoid basically all of them, though there's other issues (users pasting very long "words" to break your layout, but that's only a nuisance, not a security problem). If you do let users use HTML (for anything), you need to consider XSS and CSRF (though it need not be cross site).
Basically, read more, and always be thinking about security when you're adding new functionality that takes user input. It's really not that difficult to do, you just need to be aware of the issues. I've done almost exactly the same thing with Game!, it's not open source, but that really doesn't change any of the security issues.
This is one area where AMD is light years ahead of Intel. With Intel, you have to buy a Xeon and a server chipset to have ECC support, which basically is going to run you at least a grand or two just for the CPU and motherboard (at least if you want an i7 based Xeon). AMD on the other hand supports ECC across the board, and you just need a motherboard which supports it, which is most of them (total cost: <$500).
Most users won't wade through obvious glitches like that though to play a game. It's hard to proclaim "We don't need native ports! Wine runs almost everything great!"
See, I didn't say that. Absolutely, native ports are a thousand times better than running via Wine, but running via Wine is also a thousand times better than twiddling your thumbs (though, twiddling your thumbs is also a thousand times better than running 'doze).
Also, I think you vastly underestimate the typical 'doze user's tolerance for abuse. Not only have they been conditioned to tolerate it, they probably don't even notice it anymore.
Also, there's the fact that many games, Starcraft being a very good example, that run somewhat buggy. Patches have to be applied manually and menus in Bnet don't show at all as well as visual artifacts being quite common.
Yes, the Battle.net lobby doesn't render correctly in Starcraft, but it's still completely usable. More importantly, the actual game renders and plays flawlessly, and I've never used anything but a vanilla install of Wine to play it in the past 5+ years.
Certainly, Wine isn't perfect, but Starcraft is hardly a good example of that.
Steam's DRM is one of the least intrusive out there. I forget it's even there until some Slashdotter brings it up. Kind of diminishes your point when the evil DRM isn't even noticeable.
Quite the opposite. Just wait until you do notice the effects on everything you paid for but don't actually own.
Home Internet connections in north America (I'm in the US, but it's the same in Canada) almost never have bandwidth caps....
Hah! Oh dear god, I wish that was true. No, in Canada, none of the local phone/cable monopolies even offer an bandwidth capless plan, at any price. Furthermore, in most areas, you have no other choice than the local phone/cable monopolies, and that's if you're lucky.
I think the more likely situation is that there's a spectrum of gamers from casual to hardcore, with lots of people in between. RIght now, the people in between are quickly bored with casual games, but quickly frustrated with hardcore games. So, as the summary is saying, what gaming needs is more games like Game!, simple enough that anyone can pick them up and play without reading the manual, but with enough depth to actually keep people interested too.
The core problem is that security is only good security when it's transparent to the user. Of course, users won't buy products that appear to do nothing for them (even if they did actually work perfectly well), thus vendors are forced to produce bad software so that people will buy it.
Games are not representative of the userbase in general. These are the types of people that will install beta drivers for a 1% FPS boost even if it crashes their machine twice a day. The corporate space is the polar opposite of the gaming space.
Slashdot Mirror
Psst: RAII.
If you're a Java addict that's forced to write C++, then you'll do the stupid thing and allocate all your objects with new foo(). Needless to say, this is painfully stupid. If you're smart, you'll normally use RAII and not worry about cleaning up your objects, because they're on the stack.
PCI is nowhere close to being fast enough for USB 3, USB 2 sure, but not USB 3. Also, even a single 7200 RPM SATA hard drive can outstrip the bandwidth provided via a PCI slot nowadays. On the other hand, PCIe is a totally different story, and just about every motherboard these days includes at least a couple PCIe slots.
"However there is a significant cost difference that must be factored in when deciding 4 vs. 6 cores (at least last I checked, there was a huge price 'upgrade' for equal GHz +2 cores)" Not so: Newegg has the 6 core 2.8ghz AMD 1055T for $200, and Microcenter was recently offering $50 off a 1055T compatible AM3 motherboard when you bought a $200 1055T. The motherboards are only $55 so you're only paying $5 for the board. So essentially you pay $205 to get a 6-core CPU and a new motherboard. That's a very good price to make the leap to 6-cores.
I think the GP was talking about Intel pricing, where you're pretty much guaranteed to get fleeced. AMD is great on performance/$, but they don't have any chips that compete above the mid to mid-high performance market.
As for IE, some people would just rather not have to deal with third party crap. However, IE6? There really is no reason to run that anymore. Won't IE7 or 8 run on Windows 98?
No, IE7 and IE8 aren't supported on 98, ME or even 2000, XP and newer only.
You must be doing something wrong... Even on a huge story (like this one, 801 comments when I loaded it), Konqueror loaded and rendered all of the comments in a second or two at most. Admittedly, this machine has an i7 in it, but I can't believe that the chip in the N900 is two orders of magnitude slower.
Everybody is trying to improve their Javascript execution speed, so it's really not a big slip. Really, you can't blame them, after all, it lets you play Javascript games like Game! that much faster!
What on earth are you talking about? Vim works with every language.
So... none of them?
>>>What do you need a 3 GHz 6 core CPU for?
I don't. I watch HDTV on a 3 gigahertz single core Pentium
>>>What do you need 12G of RAM for?
I don't. I only have half-a-gig.
>>>What do you need a 3T hard drive for?
I don't. It's only 0.3 terabytes.
The questions aren't directed at anyone in particular, they're just today's equivalent of the same questions people could have asked 10 years ago... What do you need a 3 GHz Pentium for? What do you need 512M RAM for? What do you need a 300G hard drive for? Of course, the answers to those questions are blindingly obvious now, but they weren't then.
What I'm really trying to point out here is the absurdity of the typical kneejerk reaction of "Oh, there's nothing yet that requires $newtech, therefore it's stupid and nobody should buy it."
Finally, if you're going to go by the literal definition of "need", then your original question is loaded, as there's no correct answer. Nobody needs an Internet connection, a computer, anything electronic or even a home.
BTW why do you need 100 Mbit/s?
What do you need a 3 GHz 6 core CPU for? What do you need 12G of RAM for? What do you need a 3T hard drive for? These are all equally pointless questions, because regardless of the fact that you can't think of anything that would use the faster hardware, there's always countless ideas that would become practical (and widely implemented) when faster hardware is deployed.
And as for value, a good 128GB SSD is $300. For about $200 more, you can get 3 x 150GB Raptors and a $100 Adaptec SATA RAID controller, config it in RAID 5 and get comparable performance, not to mention a little redundancy. The extra initial investement will pay for itself in uptime over the long-term.
I'm sorry, but you're completely and hopelessly wrong. Spinning rust gets around 100 IOPS, maybe 200 at 15k RPM. The Intel X25-E gets around 10,000 IOPS. Assuming linear speedup (which you won't get anything close to), you'd need 100 rotational drives to come close to the performance of a single X25-E.
The only performance metric where SSDs and spinning rust are anywhere close is on linear read/write speeds. Sadly, that's of no consequence, because that workload only exists in benchmarks.
(Also, god help you if you put a database server on RAID 5... goodbye performance! RAID 10 or bust.)
SSDs absolutely blow spinning rust out of the water on a price/performance basis, even the initial models that cost a grand for <100G of storage. The only problem is that there's a lot of poor quality SSDs out there now that perform badly on random writes. If you balk at the price, you don't need the performance anyways, move along.
Except they didn't go it on their own. Apple forked KHTML into Webkit because KHTML was (and is) a substantially cleaner codebase than Gecko was (and is). Similarly, Google just forked Webkit instead of Gecko.
That's pretty much the best advise you can get — never trust the client. But as with everything in technology, there's no silver bullet. You also need to consider what you're passing back from the client, try to pass as little as possible, the less you pass back, the less you need to validate and the less chances you have to make mistakes. Also consider what form the data you're passing the data back from the client is in... validating integers is trivial, validating freeform text (especially if you're going to be displaying it somewhere on a page!) isn't.
In the case where you are displaying freeform text from the user (and you'll be hard pressed to avoid this), there are a lot of security problems you need to consider. Ultimately, if you don't let the user use HTML (ever), you can avoid basically all of them, though there's other issues (users pasting very long "words" to break your layout, but that's only a nuisance, not a security problem). If you do let users use HTML (for anything), you need to consider XSS and CSRF (though it need not be cross site).
Basically, read more, and always be thinking about security when you're adding new functionality that takes user input. It's really not that difficult to do, you just need to be aware of the issues. I've done almost exactly the same thing with Game!, it's not open source, but that really doesn't change any of the security issues.
This is one area where AMD is light years ahead of Intel. With Intel, you have to buy a Xeon and a server chipset to have ECC support, which basically is going to run you at least a grand or two just for the CPU and motherboard (at least if you want an i7 based Xeon). AMD on the other hand supports ECC across the board, and you just need a motherboard which supports it, which is most of them (total cost: <$500).
Thanks for the gouging Intel!
Most users won't wade through obvious glitches like that though to play a game. It's hard to proclaim "We don't need native ports! Wine runs almost everything great!"
See, I didn't say that. Absolutely, native ports are a thousand times better than running via Wine, but running via Wine is also a thousand times better than twiddling your thumbs (though, twiddling your thumbs is also a thousand times better than running 'doze).
Also, I think you vastly underestimate the typical 'doze user's tolerance for abuse. Not only have they been conditioned to tolerate it, they probably don't even notice it anymore.
I have not. As I mentioned, it's a completely vanilla Wine install with no Wine configuration necessary.
Yes, the Battle.net lobby doesn't render correctly in Starcraft, but it's still completely usable. More importantly, the actual game renders and plays flawlessly, and I've never used anything but a vanilla install of Wine to play it in the past 5+ years.
Certainly, Wine isn't perfect, but Starcraft is hardly a good example of that.
Steam's DRM is one of the least intrusive out there. I forget it's even there until some Slashdotter brings it up. Kind of diminishes your point when the evil DRM isn't even noticeable.
Quite the opposite. Just wait until you do notice the effects on everything you paid for but don't actually own.
Home Internet connections in north America (I'm in the US, but it's the same in Canada) almost never have bandwidth caps....
Hah! Oh dear god, I wish that was true. No, in Canada, none of the local phone/cable monopolies even offer an bandwidth capless plan, at any price. Furthermore, in most areas, you have no other choice than the local phone/cable monopolies, and that's if you're lucky.
I think the more likely situation is that there's a spectrum of gamers from casual to hardcore, with lots of people in between. RIght now, the people in between are quickly bored with casual games, but quickly frustrated with hardcore games. So, as the summary is saying, what gaming needs is more games like Game!, simple enough that anyone can pick them up and play without reading the manual, but with enough depth to actually keep people interested too.
There, fixed that for you.
Don't worry, it's still stored permanently.
The core problem is that security is only good security when it's transparent to the user. Of course, users won't buy products that appear to do nothing for them (even if they did actually work perfectly well), thus vendors are forced to produce bad software so that people will buy it.
Games are not representative of the userbase in general. These are the types of people that will install beta drivers for a 1% FPS boost even if it crashes their machine twice a day. The corporate space is the polar opposite of the gaming space.