TFA is a fun read. Too bad XML sucks. As Jerome and Philip Wadler write, "[T]he essence of XML is this: the problem it solves is not hard, and it does not solve the problem well."
Lisp had the same problem solved 40 years earlier. While a lot of people find S-expressions verbose, XML is quite a bit more verbose. Slava Akhmechet has a nice essay on the relationship between the two notations.
How did you mount it to the Aeron? Doesn't the Aeron use nonstandard armrest mounts? On a chair that uses the standard armrest mounts, the Evolution, along with the custom armrest pads it comes with, bolts on in place of the original armrest pads. In this arrangement, the failure mode you describe is not an issue.
Oh, I should have added, I hate the built-in trackpad (the buttons are stiff with too little travel, and you can't turn off tap-to-click, which I've always hated). So what I did was use double-sided foam tape to mount a Logitech Marble Mouse (which, despite the name, is a trackball) right on top of the trackpad. It fits great. Two stacks of three adhesive rubber bumpers (from any hardware store) support the other end, so it doesn't press on the space bar.
Yes, too bad the article doesn't show the Evolution chair-mount, which was the coolest. I have two of them, one >7 years old -- they seem quite sturdy to me. I think if they could have gotten the price down a little, they could have sold 10 times as many.
I just think the minimum registration fee charged by ICANN to register any domain should be much higher, like at least $10/year. If a speculator still wants to speculate at that price, let them. But now we have a situation where by posing as a registrar, they can speculate at pennies per year per domain -- which makes it economic for them to sit on vast farms of domains.
Why would this be better? Because it would be good for the Internet as a whole if domains tended to be owned by people who could get more value out of them than the squatters do.
Symbolics was basically out of business in about 1988. (A very small, as in no more than 2 full time people, company of that name existed until a year or two ago, but all they did with hardware was to maintain what had been manufactured by the original Symbolics.) 2004 is simply the year these photos were taken.
You're right that the very first models -- the LM-2 and 3600 -- were refrigerator-sized, but it wasn't long before they also started building some smaller models. The 3640 was very roughly 20"w x 30"h x 36"d, and the 3610/3620, which used gate arrays, was about 10"w x 24"h x 30"d -- this is the model pictured in the center and center-right photos on that page. Finally, there was the Ivory chip, which powered the MacIvory coprocessor card (this is what's being shown in the upper left photo) and the XL and UX series. I still have a working XL-1200; it's about the size of two Sun "pizza boxes" stacked vertically, maybe 16" x 16" x 8"h. I believe this machine was out in 1987.
(All dimensions guesstimated from memory -- figure a 20% margin of error.)
I don't see why hdd manufactureres are the ONE single exception to this long standing rule, and SI units be damned.
Ever hear of a "1.44MB" floppy? How many bytes do you suppose it holds? That's right... it's a double-sided version of a "720kB" floppy, so it really holds 1440KiB... which, perhaps inevitably, people started calling "1.44MB", even though that "MB" is the bastard child of the decimal and binary kilobytes, 1024000 bytes.
Once that monstrosity caught on, I'm afraid we were doomed.
I have no idea whether there's any chance focus fusion could work. But I do believe it has probably been a terrible mistake to have put all our eggs in the tokamak basket for all these years. When you don't know how to solve a problem, it's critical to keep exploring alternative approaches, especially if they're radically different. I would love to see substantially more funding for focus fusion, electrostatic confinement fusion, sonofusion, and even good old Pons and Fleischmann style cold fusion. The total would still be small compared to tokamak funding -- and who knows, maybe one of them would work out, or maybe we would learn something that turned out to be useful in the tokamak.
While there certainly are crackpots out there, I think we're too quick to dismiss ideas outside the mainstream, too eager to congratulate ourselves for knowing the truth already when we clearly don't know all of it. We need to cultivate more humility in the face of the mystery of the unknown.
But power plants are sized in megawatts, not kWh/year. I think that the 100GW number (rounding off) is the more useful, though there's nothing wrong with giving both numbers.
Your site just doesn't work. I'm trying to put together a substantial order, and sometimes when I add a CD to my cart, the previous contents of the cart are replaced by other items I didn't select. Sometimes they disappear entirely. Your code is crap, and that has to be at least part of your problem.
Don't forget the story about the early days of Google when the developers would occasionally receive a mysterious email message containing only a number. I've forgotten the exact number, alas, but it was always the same -- 31, let's suppose. Eventually they figured out that whenever they put more than 31 words on the Google home page, they would get a message with this number 31. I don't know if the sender was ever identified, but at least at the time, Google evidently took the message to heart.
I hope this story is still part of the company culture.
All that said, your post is reassuring. I hope you really mean the part about "if you see them when you want to see them".
Re:Reading between the lines
on
AMD NDA Scandal
·
· Score: 1
Benchmarks and volume shipments are very very different things.
Anyway, you misunderstand me. I'm a long-time AMD fanboi. I'm actually waiting to buy a Barcelona as soon as they come out. Have been since spring. (It's for an upgrade, so I don't have to have it any particular time.) I have no doubts it will be a good chip. But I'm worried about the delay... and the impending release of Penryn. The combination could spell real trouble for AMD.
Reading between the lines
on
AMD NDA Scandal
·
· Score: 2, Interesting
I think the reason this story is interesting is the hint it gives that AMD is having real trouble getting working Barcelona parts in any volume. Looks to me like they set this thing up because they either hoped to have good news, and then didn't, or because they just want to try to distract people from the Barcelona delays. Either way, seems like baaaaaad news.
We all love to blame Microsoft, and I certainly do too. But there's another party who's even more responsible, and it surprises me they never get mentioned. Who are these foul malefactors? Why, Unix gods Brian Kernighan and Dennis Ritchie, for using null-terminated strings in C. It's ridiculous on its face: a variable-size data structure with no bounds checking! Null-terminated strings -- unless used very, very carefully, which was certainly not part of the early Unix ethos -- give malicious parties hundreds of ways to crash programs, many of which can also be used to take over the process.
I know. Machines were so small back then, and null-terminated strings are soooo convenient. Still it was a disastrous engineering decision, and it's time people started saying so.
I have an interesting little stability story. I had a machine which was crashing every couple of weeks while running Linux. (I thought maybe I had a bad CPU -- I bought the chip off of eBay.) I put OpenSolaris on it, and saw this strange process called `fmd' chewing up 10% of my CPU cycles. Eventually I looked it up, and discovered that Solaris now has something called the fault manager which analyzes hardware faults, and lo and behold, I had a marginal DIMM that was getting lots of single-bit errors (I use ECC, of course). I replaced the DIMM (under warranty, thankfully) and the machine is now rock solid.
Not to disagree with what you posted, but... you're a lawyer, and you don't know how to spell "breach"? ("Breech" is a word, but it's not the same word at all.)
Slashdot Mirror
Whoops -- the authors of the linked paper should have been given as Jérôme Siméon and Philip Wadler. Sorry for the error.
TFA is a fun read. Too bad XML sucks. As Jerome and Philip Wadler write, "[T]he essence of XML is this: the problem it solves is not hard, and it does not solve the problem well."
Lisp had the same problem solved 40 years earlier. While a lot of people find S-expressions verbose, XML is quite a bit more verbose. Slava Akhmechet has a nice essay on the relationship between the two notations.
Now, if they offered a slightly thicker iPod touch with the hard drive from the 160 GB iPod classic for $500, I'd probably bite.
Me too. Even 32GB doesn't cut it.
Yesss!! Mazewar was my first video game too. I still remember that big green eye :)
How did you mount it to the Aeron? Doesn't the Aeron use nonstandard armrest mounts? On a chair that uses the standard armrest mounts, the Evolution, along with the custom armrest pads it comes with, bolts on in place of the original armrest pads. In this arrangement, the failure mode you describe is not an issue.
Oh, I should have added, I hate the built-in trackpad (the buttons are stiff with too little travel, and you can't turn off tap-to-click, which I've always hated). So what I did was use double-sided foam tape to mount a Logitech Marble Mouse (which, despite the name, is a trackball) right on top of the trackpad. It fits great. Two stacks of three adhesive rubber bumpers (from any hardware store) support the other end, so it doesn't press on the space bar.
Yes, too bad the article doesn't show the Evolution chair-mount, which was the coolest. I have two of them, one >7 years old -- they seem quite sturdy to me. I think if they could have gotten the price down a little, they could have sold 10 times as many.
Huh. I have a Tiger MPX (S2466) with dual Athlon MPs that's now over 4 years old and has never given me the slightest trouble, in 24/7 operation.
I just think the minimum registration fee charged by ICANN to register any domain should be much higher, like at least $10/year. If a speculator still wants to speculate at that price, let them. But now we have a situation where by posing as a registrar, they can speculate at pennies per year per domain -- which makes it economic for them to sit on vast farms of domains.
Why would this be better? Because it would be good for the Internet as a whole if domains tended to be owned by people who could get more value out of them than the squatters do.
Symbolics was basically out of business in about 1988. (A very small, as in no more than 2 full time people, company of that name existed until a year or two ago, but all they did with hardware was to maintain what had been manufactured by the original Symbolics.) 2004 is simply the year these photos were taken.
You're right that the very first models -- the LM-2 and 3600 -- were refrigerator-sized, but it wasn't long before they also started building some smaller models. The 3640 was very roughly 20"w x 30"h x 36"d, and the 3610/3620, which used gate arrays, was about 10"w x 24"h x 30"d -- this is the model pictured in the center and center-right photos on that page. Finally, there was the Ivory chip, which powered the MacIvory coprocessor card (this is what's being shown in the upper left photo) and the XL and UX series. I still have a working XL-1200; it's about the size of two Sun "pizza boxes" stacked vertically, maybe 16" x 16" x 8"h. I believe this machine was out in 1987.
(All dimensions guesstimated from memory -- figure a 20% margin of error.)
You're right, the "720k" floppy was already double-sided. Sorry. Doesn't change my main point, though.
1440 * 1024 bytes = 1474560 bytes = 1.47 * 10 ^ 9 bytes. Calling it a 1.44 MB floppy is wrong on both counts
My point exactly. It's neither 1.44 * 10 ^ 6 nor 1.44 * 2 ^ 20.
I don't see why hdd manufactureres are the ONE single exception to this long standing rule, and SI units be damned.
Ever hear of a "1.44MB" floppy? How many bytes do you suppose it holds? That's right... it's a double-sided version of a "720kB" floppy, so it really holds 1440KiB... which, perhaps inevitably, people started calling "1.44MB", even though that "MB" is the bastard child of the decimal and binary kilobytes, 1024000 bytes.
Once that monstrosity caught on, I'm afraid we were doomed.
I have no idea whether there's any chance focus fusion could work. But I do believe it has probably been a terrible mistake to have put all our eggs in the tokamak basket for all these years. When you don't know how to solve a problem, it's critical to keep exploring alternative approaches, especially if they're radically different. I would love to see substantially more funding for focus fusion, electrostatic confinement fusion, sonofusion, and even good old Pons and Fleischmann style cold fusion. The total would still be small compared to tokamak funding -- and who knows, maybe one of them would work out, or maybe we would learn something that turned out to be useful in the tokamak.
While there certainly are crackpots out there, I think we're too quick to dismiss ideas outside the mainstream, too eager to congratulate ourselves for knowing the truth already when we clearly don't know all of it. We need to cultivate more humility in the face of the mystery of the unknown.
But power plants are sized in megawatts, not kWh/year. I think that the 100GW number (rounding off) is the more useful, though there's nothing wrong with giving both numbers.
Your site just doesn't work. I'm trying to put together a substantial order, and sometimes when I add a CD to my cart, the previous contents of the cart are replaced by other items I didn't select. Sometimes they disappear entirely. Your code is crap, and that has to be at least part of your problem.
Okay, you win :) The only one I knew about was the Vigor... didn't know there was a 5-cyl TL.
Honda used to make one too. Extra points if you know the name of the car it went in.
Should be called the iPod Stroke.
Don't forget the story about the early days of Google when the developers would occasionally receive a mysterious email message containing only a number. I've forgotten the exact number, alas, but it was always the same -- 31, let's suppose. Eventually they figured out that whenever they put more than 31 words on the Google home page, they would get a message with this number 31. I don't know if the sender was ever identified, but at least at the time, Google evidently took the message to heart.
I hope this story is still part of the company culture.
All that said, your post is reassuring. I hope you really mean the part about "if you see them when you want to see them".
Benchmarks and volume shipments are very very different things.
Anyway, you misunderstand me. I'm a long-time AMD fanboi. I'm actually waiting to buy a Barcelona as soon as they come out. Have been since spring. (It's for an upgrade, so I don't have to have it any particular time.) I have no doubts it will be a good chip. But I'm worried about the delay... and the impending release of Penryn. The combination could spell real trouble for AMD.
I think the reason this story is interesting is the hint it gives that AMD is having real trouble getting working Barcelona parts in any volume. Looks to me like they set this thing up because they either hoped to have good news, and then didn't, or because they just want to try to distract people from the Barcelona delays. Either way, seems like baaaaaad news.
All I can say is, I hope they pull out of this.
We all love to blame Microsoft, and I certainly do too. But there's another party who's even more responsible, and it surprises me they never get mentioned. Who are these foul malefactors? Why, Unix gods Brian Kernighan and Dennis Ritchie, for using null-terminated strings in C. It's ridiculous on its face: a variable-size data structure with no bounds checking! Null-terminated strings -- unless used very, very carefully, which was certainly not part of the early Unix ethos -- give malicious parties hundreds of ways to crash programs, many of which can also be used to take over the process.
I know. Machines were so small back then, and null-terminated strings are soooo convenient. Still it was a disastrous engineering decision, and it's time people started saying so.
I have an interesting little stability story. I had a machine which was crashing every couple of weeks while running Linux. (I thought maybe I had a bad CPU -- I bought the chip off of eBay.) I put OpenSolaris on it, and saw this strange process called `fmd' chewing up 10% of my CPU cycles. Eventually I looked it up, and discovered that Solaris now has something called the fault manager which analyzes hardware faults, and lo and behold, I had a marginal DIMM that was getting lots of single-bit errors (I use ECC, of course). I replaced the DIMM (under warranty, thankfully) and the machine is now rock solid.
Not to disagree with what you posted, but... you're a lawyer, and you don't know how to spell "breach"? ("Breech" is a word, but it's not the same word at all.)