The Dave and Buster's near me only has contemporary video games. If they had Space Harrier or Stargate, my take-home pay would no longer make it all the way home!;)
I am saying that the use of pain to stop someone needs to be justified just as strongly as the use of lethal force.
Hmmm, have to disagree there. If you're housetraining a puppy, you can easily justify swatting them with a newspaper where you wouldn't want to just take them out back and shoot them.
Thing is, you're providing a gentler slope of escalation. You have an intermediate step between threatening deadly force and using it (or, more likely, you could threaten pain and use pain before you had to threaten and use deadly force).
Heh, now I want one. Hook it up to my motion sensor in the back yard and instead of just lighting up the yard, I can light up the dogs! Or course, with my luck it'll get triggered by a passing car and I'll end up torching my toolshed...
Just be prepared to bend waaaaay over when they mention price. I think even with our (significant) CA discount we were going to pay something on the far side of $5K for a single seat.
We wound up going with ER/Studio from Embarcadero Technologies. Same capabilities, and less than half the price. I use it for just what the original poster was asking about, reverse-engineering other people's databases so I can get a handle on their config. Best part is, I can reverse-engineer an SQL Server DB and have it spit me out a build script for Oracle to use when porting, woo hoo!
ErWin's a good tool, no doubt about that, but you pay for every feature you get.
Depends on what kind of math you're bad at. I had a horrible time with trig and calculus, but discrete math and geometry were a breeze. Dunno why, just is.
Dunno, I was an English major and I haven't had any problems finding development work in the last ~20 years. My code is solid and my documentation is immaculate...:)
It certainly helps that I had courses on algorithms & data structures, language design and compiler implementation, and information theory between the endless compare-and-contrast writing assignments and literary criticism classes. I wouldn't expect a "raw" liberal arts major to write code very well with just OJT (although no worse than anyone else with no formal training).
Just noticed that the Blade only has 32-bit PCI slots and uses EIDE drives. It probably WOULD have a hard time keeping up with a faster Athlon or Pentium chip.
I don't have much experience with Sun boxen (just a brief stint developing on a SPARCstation SLC), but I know that when they replaced my wife's 477MHz Alpha box with a dual-processor 800MHz Pentium///, her FEA software ran slower. Like run times of six hours suddenly stretched to almost seven. FEA software is difficult to parallelize well (it's mostly matrix multiplication), so I doubt they were seeing much help from the second processor, but clearly for math-intensive apps, an 800MHz Pentium is no match for a 477MHz Alpha.
And don't forget that Sun was one of the first vendors to have full 64-bit PCI slots, which gives appreciably more bandwidth. I could easily believe that system throughput could be on a par or even exceed that of a system equipped with a "faster" (higher MHz) CPU. It doesn't take pixie dust...
The future is epaper, with resolution and reflectiveness that is as good as or better than today's printed pages
The far future, maybe. Back in the 70's, the standard resolution for draft printing was 1200dpi. Hold a stamp or dollar bill up to a computer screen and try to imagine displaying the incredible detail that paper can hold on a device with less than 10,000dpi resolution. The article states that 100dpi e-paper is still 3-5 years away. Until they come up with something with at least 1000dpi (and with a viewable area around 20-30 square inches), I don't think it's a serious contender.
Um, same number of moving parts, only they're bigger. Two platters bolted onto a common spindle = 1 moving part. It's the same with the heads, they all ride on a common armature.
Jeez, I had no idea people had such bad luck with their drives. My main Linux box (5 year old PPro) has a 5 year old 2.5G drive, a 3 year old 6.4G drive, and a 2 year old 10G drive, and I've never had any trouble with them. Nor with the 9 year old 80M drive in my Quadra 700, although that doesn't get much use anymore...:/
You're free to say you think they suck, and you wouldn't be surprised if the company officers performed unnatural acts with rodents. Everyone (in the U.S., anyway) has the right to express an opinion, but you don't have the right to spread false information. Just watch your phrasing and you don't have a thing to worry about (well, one less thing, maybe!).
"Software Crisis". Ever notice that you don't hear that phrase too much anymore? (If you're too young to recognize that phrase, move along.) It used to mean the eternal 6- to 18-month delay to get new software written. Some large shops actually had development backlogs that stretched years. Then we got Structured Methods that made sure that we understood what we were supposed to be building. Reliability and usability went up some, but the backlog remained. Then, we got PCs, and OOP, and "Internet time", and 30G HDs for $100, and 256MB SIMMs for $100, and CPUs running in the gigaherz range and the "crisis" seems to have passed.
I, too, miss clever programming tricks, and weighing the time/space tradeoff, and seriously bumming code, but I think overall we're better off that we at least have the option of throwing more hardware at the problem, or of designing something that won't run acceptably on state-of-the-art hardware because you know that a year from now when it's released, the state of the art will have advanced to the point where your design is feasible.
ISTR there was a project to do this exact same thing (for exactly the same reasons) at DEC's Western Research Lab back in the mid-80s. They virtualized the VAX CPU and ran multiple copies of VMS inside a 'supervisor' version of VMS.
Weird commenting style, though -- he documents the hell out of his algorithms and data structures, but doesn't comment his code that much (no mention made of function return values or error conditions raised), and the thing I thought was weirdest -- he has plenty of #ifdef cplusplus statements, but then he goes ahead and uses '//' comments!
Not that mymalloc's any great shakes (e-mail for source)...
One neat use I'd like to see would be in labs for O/S classes. Recompile your kernel, download it to your other box and run it. If it crashes, no problem -- it won't interfere with your development/monitoring box. And it might make it possible to do a "remote" crash analysis for debugging.
Actually, it had a major part in the non-free software movement, too. Legend has it Paul Allen write an 8080 simulator on the PDP-10, and that's what he and Bill Gates used to write BASIC for the Altair.
Didn't the PDP-10 have some sort of an asychronous backplane or something? ISTR that there was much twitchiness involved in installing new cards & such.
Re:how fast were these things?
on
PDP-10 Revival
·
· Score: 1
An quick list of various PDP-10 capabilities is here
Re:how fast were these things?
on
PDP-10 Revival
·
· Score: 1
Some of them were actually variable-speed. There was a big knob on the console so you could vary the system clock. I only heard of field circus using it when running diagnostics.
Compiled-in? I don't think so. That account ships as one of the default accounts (along with SYSTEM/manager), and exists so that the installation crew can do their job. Field circus accts should always be disabled until the CSE arrives, then re-authorized with a random password for the duration of the maintenance, and then changed as soon as the CSE leaves the premises.
Heh. And let's not forget "slipstreaming", the entertaining practice of providing updated code in DLLs and not changing the version number. We once had some code that wouldn't work on a user's machine because he hadn't installed the latest version of Excel, even though he had the right version of the DLLs. And of course the "improved" version was only a couple hundred bytes larger, so since Windows Explorer displays file sizes to the nearest K, they looked identical. Not that I'm frustrated and bitter, mind you...
Slashdot Mirror
The Dave and Buster's near me only has contemporary video games. If they had Space Harrier or Stargate, my take-home pay would no longer make it all the way home! ;)
I am saying that the use of pain to stop someone needs to be justified just as strongly as the use of lethal force.
Hmmm, have to disagree there. If you're housetraining a puppy, you can easily justify swatting them with a newspaper where you wouldn't want to just take them out back and shoot them.
Thing is, you're providing a gentler slope of escalation. You have an intermediate step between threatening deadly force and using it (or, more likely, you could threaten pain and use pain before you had to threaten and use deadly force).
Heh, now I want one. Hook it up to my motion sensor in the back yard and instead of just lighting up the yard, I can light up the dogs! Or course, with my luck it'll get triggered by a passing car and I'll end up torching my toolshed...
Just be prepared to bend waaaaay over when they mention price. I think even with our (significant) CA discount we were going to pay something on the far side of $5K for a single seat.
We wound up going with ER/Studio from Embarcadero Technologies. Same capabilities, and less than half the price. I use it for just what the original poster was asking about, reverse-engineering other people's databases so I can get a handle on their config. Best part is, I can reverse-engineer an SQL Server DB and have it spit me out a build script for Oracle to use when porting, woo hoo!
ErWin's a good tool, no doubt about that, but you pay for every feature you get.
Depends on what kind of math you're bad at. I had a horrible time with trig and calculus, but discrete math and geometry were a breeze. Dunno why, just is.
Dunno, I was an English major and I haven't had any problems finding development work in the last ~20 years. My code is solid and my documentation is immaculate... :)
It certainly helps that I had courses on algorithms & data structures, language design and compiler implementation, and information theory between the endless compare-and-contrast writing assignments and literary criticism classes. I wouldn't expect a "raw" liberal arts major to write code very well with just OJT (although no worse than anyone else with no formal training).
Just noticed that the Blade only has 32-bit PCI slots and uses EIDE drives. It probably WOULD have a hard time keeping up with a faster Athlon or Pentium chip.
I don't have much experience with Sun boxen (just a brief stint developing on a SPARCstation SLC), but I know that when they replaced my wife's 477MHz Alpha box with a dual-processor 800MHz Pentium ///, her FEA software ran slower. Like run times of six hours suddenly stretched to almost seven. FEA software is difficult to parallelize well (it's mostly matrix multiplication), so I doubt they were seeing much help from the second processor, but clearly for math-intensive apps, an 800MHz Pentium is no match for a 477MHz Alpha.
And don't forget that Sun was one of the first vendors to have full 64-bit PCI slots, which gives appreciably more bandwidth. I could easily believe that system throughput could be on a par or even exceed that of a system equipped with a "faster" (higher MHz) CPU. It doesn't take pixie dust...
The future is epaper, with resolution and reflectiveness that is as good as or better than today's printed pages
The far future, maybe. Back in the 70's, the standard resolution for draft printing was 1200dpi. Hold a stamp or dollar bill up to a computer screen and try to imagine displaying the incredible detail that paper can hold on a device with less than 10,000dpi resolution. The article states that 100dpi e-paper is still 3-5 years away. Until they come up with something with at least 1000dpi (and with a viewable area around 20-30 square inches), I don't think it's a serious contender.
more platters = more moving parts
Um, same number of moving parts, only they're bigger. Two platters bolted onto a common spindle = 1 moving part. It's the same with the heads, they all ride on a common armature.
Jeez, I had no idea people had such bad luck with their drives. My main Linux box (5 year old PPro) has a 5 year old 2.5G drive, a 3 year old 6.4G drive, and a 2 year old 10G drive, and I've never had any trouble with them. Nor with the 9 year old 80M drive in my Quadra 700, although that doesn't get much use anymore... :/
As long as you don't commit libel, anyway.
You're free to say you think they suck, and you wouldn't be surprised if the company officers performed unnatural acts with rodents. Everyone (in the U.S., anyway) has the right to express an opinion, but you don't have the right to spread false information. Just watch your phrasing and you don't have a thing to worry about (well, one less thing, maybe!).
"Software Crisis". Ever notice that you don't hear that phrase too much anymore? (If you're too young to recognize that phrase, move along.) It used to mean the eternal 6- to 18-month delay to get new software written. Some large shops actually had development backlogs that stretched years. Then we got Structured Methods that made sure that we understood what we were supposed to be building. Reliability and usability went up some, but the backlog remained. Then, we got PCs, and OOP, and "Internet time", and 30G HDs for $100, and 256MB SIMMs for $100, and CPUs running in the gigaherz range and the "crisis" seems to have passed.
I, too, miss clever programming tricks, and weighing the time/space tradeoff, and seriously bumming code, but I think overall we're better off that we at least have the option of throwing more hardware at the problem, or of designing something that won't run acceptably on state-of-the-art hardware because you know that a year from now when it's released, the state of the art will have advanced to the point where your design is feasible.
ISTR there was a project to do this exact same thing (for exactly the same reasons) at DEC's Western Research Lab back in the mid-80s. They virtualized the VAX CPU and ran multiple copies of VMS inside a 'supervisor' version of VMS.
MAN, that's some well-commented code!
Weird commenting style, though -- he documents the hell out of his algorithms and data structures, but doesn't comment his code that much (no mention made of function return values or error conditions raised), and the thing I thought was weirdest -- he has plenty of #ifdef cplusplus statements, but then he goes ahead and uses '//' comments!
Not that my malloc's any great shakes (e-mail for source)...
Speak for yourself, meat person
Frankly I prefer the term "natural person" myself...
One neat use I'd like to see would be in labs for O/S classes. Recompile your kernel, download it to your other box and run it. If it crashes, no problem -- it won't interfere with your development/monitoring box. And it might make it possible to do a "remote" crash analysis for debugging.
You don't use continue or break statements anywhere? Just because they don't have an explicit label doesn't mean they're not GOTOs...
Actually, it had a major part in the non-free software movement, too. Legend has it Paul Allen write an 8080 simulator on the PDP-10, and that's what he and Bill Gates used to write BASIC for the Altair.
Didn't the PDP-10 have some sort of an asychronous backplane or something? ISTR that there was much twitchiness involved in installing new cards & such.
An quick list of various PDP-10 capabilities is here
Some of them were actually variable-speed. There was a big knob on the console so you could vary the system clock. I only heard of field circus using it when running diagnostics.
User: Field Password : Engineer
Compiled-in? I don't think so. That account ships as one of the default accounts (along with SYSTEM/manager), and exists so that the installation crew can do their job. Field circus accts should always be disabled until the CSE arrives, then re-authorized with a random password for the duration of the maintenance, and then changed as soon as the CSE leaves the premises.
If you don't mind lots of false alarms after that big chili dinner...
Heh. And let's not forget "slipstreaming", the entertaining practice of providing updated code in DLLs and not changing the version number. We once had some code that wouldn't work on a user's machine because he hadn't installed the latest version of Excel, even though he had the right version of the DLLs. And of course the "improved" version was only a couple hundred bytes larger, so since Windows Explorer displays file sizes to the nearest K, they looked identical. Not that I'm frustrated and bitter, mind you...