Right, but downloading, say, 50 GB at, say 3 MB/s (reasonable rate for most people on campuses (campi?)) only takes 5 hours... sure, it's not quite streaming, but it takes half as long to watch as to download.
They do back it up, for 12 months, or, if you want to buy the premium warrantee, for 36 months. You certainly shouldn't trust any unwritten 'promise,' and, in fact, I never mentioned a promise. But I do wonder whether the people who say "what should I do?" ever considered just being polite and asking for help.
What company/doesn't/ have a quality/money tradeoff? Features are, in today's electrical age, essentially free, up to a certain point. Even when you have a feature tradeoff, that can only take you so far... and on the cheaper item, the consumer really expects it to be cheap, so costs must be cut somehow. Look at Apple, yes, with there 12" iBook vs. 12" pBook. Look at Canon and there consumer vs. L series lenses. Look at any bike company, say Shimano. Look at movie theatres. Look at shoe companies!
What are they going to do when the warranty runs out?
Do what I do. Ask nicely if they'll fix it anyway. In every case Apple has said yes, and a few times has specifically said "if it's been less than one year since your last repair, we're authorized to do it." On the other hand, when my Sony died, no luck. When my Dell died, no luck. When my Compaq died... well, honestly, I was so happy about that thing dying that I didn't ask.
Gateway might be different... like Apple, they actually have a place you can go to ask nicely. Interaction in person is much more powerful than interaction over the phone.
A little of both. The iBook is the cheapest, lowest-margin laptop Apple has ever made, and it has had a high failure rate. On the other hand, even if the failure rate is half that of Dell, say, you'll here more about it because Apple users expect more. So the iBook does suck, quality-wise, for an Apple (I'm typing from one now), but it's still better than a PC. Also, Apple has been fixing the iBooks out of warrantee, if you ask nicely. I can't imagine Sony ever doing that.
OS X and all Apple software for OS X that I know of (maybe not including some high-end stuff, I just don't know for sure) is indeed built with GCC.
Porting an operating system to a different compiler is a pain in the neck, and most OS's use compiler-specific tricks to deal with low-level details. Also, most of Apple's higher-level software is in Objective C, and as of now only GCC really supports Obj C well on the mac.
Personally, I just (automatically) throw out any e-mail with more than two misspelled words. As long as people don't use middle names, and aren't idiots, it works out.
I'd tend to agree with that (can't tell if you're being sarcastic or not)... but then again, I use laptops and mini-itx boards for 90% of what I do. Processor life cycles are only about a factor of two shorter than supporting technology life cycles; don't charge me half the cost of another processor just so I can spend more money on another processor which will be limited by the motherboard.
Okay... but what about when I come up with a chip that uses 500W per package, uses quad 256-bit busses, uses 0.7V instead of 1.3V (higher amperage means even less resistance allowed)?
Okay, fine, we saw that far ahead, so our generic package had 5000 pins (about what you'd need for the above, judging by the Power series).
Let's ignore that that package will cost 10x more than current cpu packages, and that the motherboard will cost $10 more, too. Let's explain that away by "economy of scale."
Then I come up with a chip that uses only 20W and uses a high-speed serial 16-bit memory interface. Now, let's go with the "cost is no issue" argument... we want to cram it into our $50 socket, instead of making a $5 socket that's adequate. But wait! Our socket was designed with a 500Mhz FSB in mind, now we need to run it at 3GHz for that memory bus... and our pins' electrical connection wasn't rated for that speed!
Do you propose to design a socket for every possible chip in the next five (say) years now? Do you claim to know what technology we'll be using in five years?
Do you think a 10000 pin chip would be faster than a 4000 pin chip or something?
Not to put too fine a point on it, yes. Let's pretend there are only two things going over all those pins: access to main memory, and power.
How many power pins do we need? On modern (Pentium and further) processors, a good guideline is that half of the pins are power pins, although sometimes its as low as a third. Todays processors use at least 400 power pins. Why so many? Because it takes really low resistance to get 60A across a wire effectively... this includes the pins themselves, but also inside the chip the distance power can travel is limitted. Can you reduce this number of pins? Well, yes, if you're willing to reduce the quality of your power supply... which means decreasing the clock speed to keep reliability.
How many memory pins do we need? It depends. If we just want to be able to read and write a 32-bit bus, we can probably pull it off "efficiently" with 33 pins or so. We can use even less pins, of course, if we take multiple bus cycles to read and write memory. But what if we want a wider memory bus? Todays systems use 64 and 128 bit memory busses... many of tomorrows will use 256 bit ones. This takes pins, honestly. A 486 package just can't support a 128 bit memory bus. Can you reduce this number of pins? Well, yes. Either reduce the bus width, or use a narrow serial bus width (rambus style) and do and on-chip memory controller.
Not even getting into IO... why do you think they keep increasing pin count? Do you think a 10000 pin chip can really be trimmed to a 4000 pin chip without making huge compromises?
This is a wretched idea. Lets assume for a moment you could do the backwards compatibility, which means changing what pins mean, changing bus widths, and changing voltages... you still run into the problem that pin numbers only go up (from 32 on my first computer to over 4000 on the Power4 I play with now)... which means you'd have to supply more pins than you need. Even if you only supply twice as many pins as needed today... the socket is the most expensive component on most motherboards, and making it more expensive would be deeply silly.
Um, no. The force of the impact is equal to the mass of the object times the negative deceleration of the object as it collides and it and the floor compress. The acceleration of gravity has nothing to do with it, except in determining the speed from which the object will decelerate.
Okay, I give up. How do you make your dirt do logs? Mine was strictly five-function. Addition, subtraction, multiplication, division are straightforward... for square root, you make a flattened pile, and throw a piece of chewing gum at it. Then, count the number of grains across the diameter of the crater. And then chew the gum... if you allow the dirt to accumulate, you'll get errors!
Dragon Warrior IV is the first game I can think of that had real adaptive AI, but it was AI on your side... the bad critters never changed but, in Chapter 5, your team members would keep track of "what worked" (seems to just be how close an action was to a kill), and increase the probability of choosing that action. It was a good implementation, too. They started with 5 (I think) user-selectable probability distributions (defensive, offensive, etc), and when a given one was selected, training would only happen within it... so you essentially have five modes within which learning takes place, separately, and the user could (and occasionally really should) disable those modes which learned weird habits.
Ugh. There used to be a law in the US that speedometers had to top off at 85 mph. My car ('93 Explorer) has that, and it's damnably unsafe. Most of the time, I have no fucking clue how fast I'm going, other than "as fast as the cars around me." I've taken to using a GPS on I-280, just to know if I'm going under 100 (at which point the cops might care, since I'm obstructing traffic going so slow) or not.
Want my honest advice? Next time there's even the slightest problem (even just software), go to the nearest apple store, ask them to pull up your repair record, and say (preferably in an angry tone of voice) "this number of repairs is unacceptable, at what point do you do a replacement?" Now, they can't answer that directly... but each component has a certain "score", and if you exceed a certain number of points of replacements, they are allowed to (NOT required to) replace the whole machine. Go talk to a human, in person, and make it clear it's what you want.
One warning: The new machine comes with a one year warrantee, NOT the remainder of your AppleCare. If you want, you can call up AppleCare and yell at them about this once you get your new machine. My friend who did this got his AppleCare transfered, but it was an exception, not a rule.
Again, if you've had that many repairs, you qualify for a new machine, but that doesn't mean they'll give you one unless (a) you ask or (b, my approach) you're pissed enough that they offer just to shut you up.
Slashdot Mirror
Right, but downloading, say, 50 GB at, say 3 MB/s (reasonable rate for most people on campuses (campi?)) only takes 5 hours... sure, it's not quite streaming, but it takes half as long to watch as to download.
They do back it up, for 12 months, or, if you want to buy the premium warrantee, for 36 months. You certainly shouldn't trust any unwritten 'promise,' and, in fact, I never mentioned a promise. But I do wonder whether the people who say "what should I do?" ever considered just being polite and asking for help.
What company /doesn't/ have a quality/money tradeoff? Features are, in today's electrical age, essentially free, up to a certain point. Even when you have a feature tradeoff, that can only take you so far... and on the cheaper item, the consumer really expects it to be cheap, so costs must be cut somehow. Look at Apple, yes, with there 12" iBook vs. 12" pBook. Look at Canon and there consumer vs. L series lenses. Look at any bike company, say Shimano. Look at movie theatres. Look at shoe companies!
What are they going to do when the warranty runs out?
Do what I do. Ask nicely if they'll fix it anyway. In every case Apple has said yes, and a few times has specifically said "if it's been less than one year since your last repair, we're authorized to do it." On the other hand, when my Sony died, no luck. When my Dell died, no luck. When my Compaq died... well, honestly, I was so happy about that thing dying that I didn't ask.
Gateway might be different... like Apple, they actually have a place you can go to ask nicely. Interaction in person is much more powerful than interaction over the phone.
A little of both. The iBook is the cheapest, lowest-margin laptop Apple has ever made, and it has had a high failure rate. On the other hand, even if the failure rate is half that of Dell, say, you'll here more about it because Apple users expect more. So the iBook does suck, quality-wise, for an Apple (I'm typing from one now), but it's still better than a PC. Also, Apple has been fixing the iBooks out of warrantee, if you ask nicely. I can't imagine Sony ever doing that.
Downside: Striping must be done in software, with a CPU hit. Upside: You can RAID 5 instead of stripe, if you're not totally insane.
Are you going to trust google's cache, or google?
Whoops. And the reason that the math doesn't agree with yours is that (in my opinion correctly) Google says 1TB = 1TiB.
Or, to save the math, and assist google in taking over the world: link.
OS X and all Apple software for OS X that I know of (maybe not including some high-end stuff, I just don't know for sure) is indeed built with GCC.
Porting an operating system to a different compiler is a pain in the neck, and most OS's use compiler-specific tricks to deal with low-level details. Also, most of Apple's higher-level software is in Objective C, and as of now only GCC really supports Obj C well on the mac.
Personally, I just (automatically) throw out any e-mail with more than two misspelled words. As long as people don't use middle names, and aren't idiots, it works out.
I'd tend to agree with that (can't tell if you're being sarcastic or not)... but then again, I use laptops and mini-itx boards for 90% of what I do. Processor life cycles are only about a factor of two shorter than supporting technology life cycles; don't charge me half the cost of another processor just so I can spend more money on another processor which will be limited by the motherboard.
Okay... but what about when I come up with a chip that uses 500W per package, uses quad 256-bit busses, uses 0.7V instead of 1.3V (higher amperage means even less resistance allowed)?
Okay, fine, we saw that far ahead, so our generic package had 5000 pins (about what you'd need for the above, judging by the Power series).
Let's ignore that that package will cost 10x more than current cpu packages, and that the motherboard will cost $10 more, too. Let's explain that away by "economy of scale."
Then I come up with a chip that uses only 20W and uses a high-speed serial 16-bit memory interface. Now, let's go with the "cost is no issue" argument... we want to cram it into our $50 socket, instead of making a $5 socket that's adequate. But wait! Our socket was designed with a 500Mhz FSB in mind, now we need to run it at 3GHz for that memory bus... and our pins' electrical connection wasn't rated for that speed!
Do you propose to design a socket for every possible chip in the next five (say) years now? Do you claim to know what technology we'll be using in five years?
Do you think a 10000 pin chip would be faster than a 4000 pin chip or something?
Not to put too fine a point on it, yes. Let's pretend there are only two things going over all those pins: access to main memory, and power.
How many power pins do we need? On modern (Pentium and further) processors, a good guideline is that half of the pins are power pins, although sometimes its as low as a third. Todays processors use at least 400 power pins. Why so many? Because it takes really low resistance to get 60A across a wire effectively... this includes the pins themselves, but also inside the chip the distance power can travel is limitted. Can you reduce this number of pins? Well, yes, if you're willing to reduce the quality of your power supply... which means decreasing the clock speed to keep reliability.
How many memory pins do we need? It depends. If we just want to be able to read and write a 32-bit bus, we can probably pull it off "efficiently" with 33 pins or so. We can use even less pins, of course, if we take multiple bus cycles to read and write memory. But what if we want a wider memory bus? Todays systems use 64 and 128 bit memory busses... many of tomorrows will use 256 bit ones. This takes pins, honestly. A 486 package just can't support a 128 bit memory bus. Can you reduce this number of pins? Well, yes. Either reduce the bus width, or use a narrow serial bus width (rambus style) and do and on-chip memory controller.
Not even getting into IO... why do you think they keep increasing pin count? Do you think a 10000 pin chip can really be trimmed to a 4000 pin chip without making huge compromises?
This is a wretched idea. Lets assume for a moment you could do the backwards compatibility, which means changing what pins mean, changing bus widths, and changing voltages... you still run into the problem that pin numbers only go up (from 32 on my first computer to over 4000 on the Power4 I play with now)... which means you'd have to supply more pins than you need. Even if you only supply twice as many pins as needed today... the socket is the most expensive component on most motherboards, and making it more expensive would be deeply silly.
I sympathize. I used to have a '68 volvo with a built-in shower.
It's not a problem, it's a biological coupon!
Um, no. The force of the impact is equal to the mass of the object times the negative deceleration of the object as it collides and it and the floor compress. The acceleration of gravity has nothing to do with it, except in determining the speed from which the object will decelerate.
Okay, I give up. How do you make your dirt do logs? Mine was strictly five-function. Addition, subtraction, multiplication, division are straightforward... for square root, you make a flattened pile, and throw a piece of chewing gum at it. Then, count the number of grains across the diameter of the crater. And then chew the gum... if you allow the dirt to accumulate, you'll get errors!
See my comment here for an old, old case that I do consider learning.
Dragon Warrior IV is the first game I can think of that had real adaptive AI, but it was AI on your side... the bad critters never changed but, in Chapter 5, your team members would keep track of "what worked" (seems to just be how close an action was to a kill), and increase the probability of choosing that action. It was a good implementation, too. They started with 5 (I think) user-selectable probability distributions (defensive, offensive, etc), and when a given one was selected, training would only happen within it... so you essentially have five modes within which learning takes place, separately, and the user could (and occasionally really should) disable those modes which learned weird habits.
Phrases to avoid in the future: "despite what you think, do you think..."
Ugh. There used to be a law in the US that speedometers had to top off at 85 mph. My car ('93 Explorer) has that, and it's damnably unsafe. Most of the time, I have no fucking clue how fast I'm going, other than "as fast as the cars around me." I've taken to using a GPS on I-280, just to know if I'm going under 100 (at which point the cops might care, since I'm obstructing traffic going so slow) or not.
Well, if I got an e-mail with that sig, I would treat it as junk, regardless of the contents. Works for me.
Normal warrantee.
Want my honest advice? Next time there's even the slightest problem (even just software), go to the nearest apple store, ask them to pull up your repair record, and say (preferably in an angry tone of voice) "this number of repairs is unacceptable, at what point do you do a replacement?" Now, they can't answer that directly... but each component has a certain "score", and if you exceed a certain number of points of replacements, they are allowed to (NOT required to) replace the whole machine. Go talk to a human, in person, and make it clear it's what you want.
One warning: The new machine comes with a one year warrantee, NOT the remainder of your AppleCare. If you want, you can call up AppleCare and yell at them about this once you get your new machine. My friend who did this got his AppleCare transfered, but it was an exception, not a rule.
Again, if you've had that many repairs, you qualify for a new machine, but that doesn't mean they'll give you one unless (a) you ask or (b, my approach) you're pissed enough that they offer just to shut you up.