My son builds the set once (or maybe I build it for him), and then after it's destroyed in the a lego battle he builds whatever he wants with it. Occasionally he will want to rebuild a set, but half his (considerable) lego are mixed in a large bin so it can be a tedious process!
I think its more about how parents and society encourage kids to use the sets than what they actually start out as.
Last Gasp *was depressing* I read it in the late 80s and was down for weeks. It was so realistic I found myself checking the newspaper and TV news for updates... Good choice!
I'd have to go for "The Ugly Little Boy" by Asimov (http://en.wikipedia.org/wiki/The_Ugly_Little_Boy). It's sad in an unexpected way for a Sci-Fi story that's not easily forgotten.
Brave New World and Fahrenheit 451 are also good choices, but I wanted to add something unique.
Seriously... they can't even get the facts straight. You're right this is NOT a disney film, it's Universal. I'm a disney geek I'd know, but IMDB will back us up. Oh, and they've butchered the story, but that's another story.
My wife and & I "share" our cars... which arises from the fact that we only have one child seat. So whoever has our son at the moment, gets our Toyota Sienna, and whoever doesn't, gets our Toyata Camry. Both of these vehicles have a 1MZFE toyata v6, the sienna with the vvti head and the camry without, which translates into a modest HP increase for the sienna (about 25hp) the torque ratings are the same. The sienna is about 300 lb heavier than the camry. Both have automatics, and I believe both are EPA rated 19/26.
I've always tracked my mileage by going from full tank to full tank, but lately I found a palm app called "Fuel Log" (Open Source Free SW = cool) and have been using this for the last several months to track our milage.
This is where things get odd... The sienna over the same city/suburban driving routes with the same driver consistently gets better gas milage than the camry, like 21 vs 19. But on the highway, the camry does better, usually around 27, vs 24-26 for the sienna. When I drive a specific car it always gets better gas milage than when my wife drives it. I tend to drive harder than my wife does... go figure
Our previous vehicles usually did much better than EPA ratings. My old Escort GT was rated 25/30 I believe, and I never once got below 30, usually averaged around 32, and got as high as 35 on a couple of long hwy trips. My wife's old Mercury Tracer would consistently beat it's 29/35 EPA rating, and on a couple of long trips even topped 40mpg... Our even more previous car (94 honda accord) was much the same way.
Leads me to wonder if Toyotas simply don't match well to the EPA test. Who does these tests anyway? The manufacturer or the EPA or some 3rd party? Anybody Know???
Keep in mind that EPA cycles probably make the assumption that you stick to the speed limits:) People here in Atlanta seem to treat the roads as a raceway... average freeway speeds are probably up around 75-80MPH even where the legal limit is 55MPH. That's gotta have an effect.
I would have to say Light Weight Processes (LWPs) are not only missing but highly desired.
In Linux a thread of execution maps to a process, and is scheduled as such. (that's how I understand it anyway). This is fine and dandy if you're looking for a quick way to implement threads without a wholesale rewrite of the scheduler and process structures, but for context switch times between threads it leaves a bit to be desired.
In Solaris for example a thread maps to a LWP, which is seperate from a "heavyweight" process (for lack of a better term). A given process can have many threads, and therefore many LWPs in solaris. The advantage here is that the scheduler deals with LWPs and not just a whole process. Why is this an advantage?
Because when the scheduler moves from one LWP of a process to another on solaris it doesn't incur a full context switch. It simply jumps to a new point of execution in the code and moves on, saving all the time necessary to push all of the registers and so on onto the stack.
In my experience HEAVILY threaded apps (100+ threads in the same process, can we say java here?) perform far better under solaris than linux on the same hardware. In the code I tested it was TWICE as fast on solaris.So just maybe LWPs and a scheduler capable of dealing with them would be a nice thing to have.
Being a former embedded systems software engineer who did a fair amount of work in the modeling & simulation environment, I've seen a lot of engineering problems "solved" through code. And let me tell you that you're right, few if any are done in an OOP language. Why?
Perforance, plain & simple. OOP languages (I'm simply going to use C++ in place of OOP as it's easier for me to type:) I'm not trying to imply that C++ should stand as a definition for all OOP languages) are not deterministic, and don't typically run too efficently.
There are several reasons for this, but two stand out in my mind. Memory Management, and what I'll call Bloat.
First off, C++ code tends to make use of a LOT of dynamically allocated memory. Lots of structures passed back and forth between procedures (methods), things on the heap, and so on. All of this adds overhead to the code. It takes time to do all of that. Just puttng a structure on the stack can result in thousnads of ops (read mem, write mem...). THen there's garbage collection (GC), this runs async to everything else and can eat up more CPU cycles you process needs to run. it may also make some sections of code run longer than others because the code has to wait for the GC to finish. Enough of that.
Bloat. In order to meet one goal (quicker code development, and lots of reuse of code) C++ has added many layers on top of C. function tables, polymorphism, abstraction, overloading... these things all add more execution time to the code, and make it less deterministic. When I add 2 variables in C or FORTRAN, I know about how long that takes. if I do it in C++ with abstract datatypes, who knows? The complier has to generate logic that can to figure out what function to use, call it (copy the data onto the stack) perform the operation, and return the result. That logic all takes time to run.
For solving a complicated problem in a timely manner, like say modeling the amount of radar energy reflected off the complex surface of an aircraft, and returning that result to a flight sim C++ might take too long. By the time the code figures out if you can see the other guy, he's already sent a missile your way!
More complicated problems, like modeling a weather pattern, FEA, or hydrodynamics caan take very long times to run without all of the overhead of C++. The also tend to be fairly straight forward problems to solve, just very compute intensive. Many business apps on the other hand (which is what C++ is good at) have a myriad of nuances and abstractions that aren't comput intensive, but require lots of logic and code to figure out which path to take. Like generating the financial statements for a company. it may only be a list of numbers to add up, but where to find the numbers? what do they mean? What about depreciation, partially owned assets, and one time charges. Lots to figure out what to do with, but little calculation to be done.
Slashdot Mirror
My son builds the set once (or maybe I build it for him), and then after it's destroyed in the a lego battle he builds whatever he wants with it. Occasionally he will want to rebuild a set, but half his (considerable) lego are mixed in a large bin so it can be a tedious process! I think its more about how parents and society encourage kids to use the sets than what they actually start out as.
Last Gasp *was depressing* I read it in the late 80s and was down for weeks. It was so realistic I found myself checking the newspaper and TV news for updates... Good choice!
I'd have to go for "The Ugly Little Boy" by Asimov (http://en.wikipedia.org/wiki/The_Ugly_Little_Boy). It's sad in an unexpected way for a Sci-Fi story that's not easily forgotten. Brave New World and Fahrenheit 451 are also good choices, but I wanted to add something unique.
Seriously... they can't even get the facts straight. You're right this is NOT a disney film, it's Universal. I'm a disney geek I'd know, but IMDB will back us up. Oh, and they've butchered the story, but that's another story.
My wife and & I "share" our cars... which arises from the fact that we only have one child seat. So whoever has our son at the moment, gets our Toyota Sienna, and whoever doesn't, gets our Toyata Camry. Both of these vehicles have a 1MZFE toyata v6, the sienna with the vvti head and the camry without, which translates into a modest HP increase for the sienna (about 25hp) the torque ratings are the same. The sienna is about 300 lb heavier than the camry. Both have automatics, and I believe both are EPA rated 19/26.
:) People here in Atlanta seem to treat the roads as a raceway... average freeway speeds are probably up around 75-80MPH even where the legal limit is 55MPH. That's gotta have an effect.
I've always tracked my mileage by going from full tank to full tank, but lately I found a palm app called "Fuel Log" (Open Source Free SW = cool) and have been using this for the last several months to track our milage.
This is where things get odd... The sienna over the same city/suburban driving routes with the same driver consistently gets better gas milage than the camry, like 21 vs 19. But on the highway, the camry does better, usually around 27, vs 24-26 for the sienna. When I drive a specific car it always gets better gas milage than when my wife drives it. I tend to drive harder than my wife does... go figure
Our previous vehicles usually did much better than EPA ratings. My old Escort GT was rated 25/30 I believe, and I never once got below 30, usually averaged around 32, and got as high as 35 on a couple of long hwy trips. My wife's old Mercury Tracer would consistently beat it's 29/35 EPA rating, and on a couple of long trips even topped 40mpg... Our even more previous car (94 honda accord) was much the same way.
Leads me to wonder if Toyotas simply don't match well to the EPA test. Who does these tests anyway? The manufacturer or the EPA or some 3rd party? Anybody Know???
Keep in mind that EPA cycles probably make the assumption that you stick to the speed limits
Heaven forbid! A serious suggestion!
I would have to say Light Weight Processes (LWPs) are not only missing but highly desired.
In Linux a thread of execution maps to a process, and is scheduled as such. (that's how I understand it anyway). This is fine and dandy if you're looking for a quick way to implement threads without a wholesale rewrite of the scheduler and process structures, but for context switch times between threads it leaves a bit to be desired.
In Solaris for example a thread maps to a LWP, which is seperate from a "heavyweight" process (for lack of a better term). A given process can have many threads, and therefore many LWPs in solaris. The advantage here is that the scheduler deals with LWPs and not just a whole process. Why is this an advantage?
Because when the scheduler moves from one LWP of a process to another on solaris it doesn't incur a full context switch. It simply jumps to a new point of execution in the code and moves on, saving all the time necessary to push all of the registers and so on onto the stack.
In my experience HEAVILY threaded apps (100+ threads in the same process, can we say java here?) perform far better under solaris than linux on the same hardware. In the code I tested it was TWICE as fast on solaris.So just maybe LWPs and a scheduler capable of dealing with them would be a nice thing to have.
Just some food for thought.
Being a former embedded systems software engineer who did a fair amount of work in the modeling & simulation environment, I've seen a lot of engineering problems "solved" through code. And let me tell you that you're right, few if any are done in an OOP language. Why?
:) I'm not trying to imply that C++ should stand as a definition for all OOP languages) are not deterministic, and don't typically run too efficently.
Perforance, plain & simple. OOP languages (I'm simply going to use C++ in place of OOP as it's easier for me to type
There are several reasons for this, but two stand out in my mind. Memory Management, and what I'll call Bloat.
First off, C++ code tends to make use of a LOT of dynamically allocated memory. Lots of structures passed back and forth between procedures (methods), things on the heap, and so on. All of this adds overhead to the code. It takes time to do all of that. Just puttng a structure on the stack can result in thousnads of ops (read mem, write mem...). THen there's garbage collection (GC), this runs async to everything else and can eat up more CPU cycles you process needs to run. it may also make some sections of code run longer than others because the code has to wait for the GC to finish. Enough of that.
Bloat. In order to meet one goal (quicker code development, and lots of reuse of code) C++ has added many layers on top of C. function tables, polymorphism, abstraction, overloading... these things all add more execution time to the code, and make it less deterministic. When I add 2 variables in C or FORTRAN, I know about how long that takes. if I do it in C++ with abstract datatypes, who knows? The complier has to generate logic that can to figure out what function to use, call it (copy the data onto the stack) perform the operation, and return the result. That logic all takes time to run.
For solving a complicated problem in a timely manner, like say modeling the amount of radar energy reflected off the complex surface of an aircraft, and returning that result to a flight sim C++ might take too long. By the time the code figures out if you can see the other guy, he's already sent a missile your way!
More complicated problems, like modeling a weather pattern, FEA, or hydrodynamics caan take very long times to run without all of the overhead of C++. The also tend to be fairly straight forward problems to solve, just very compute intensive. Many business apps on the other hand (which is what C++ is good at) have a myriad of nuances and abstractions that aren't comput intensive, but require lots of logic and code to figure out which path to take. Like generating the financial statements for a company. it may only be a list of numbers to add up, but where to find the numbers? what do they mean? What about depreciation, partially owned assets, and one time charges. Lots to figure out what to do with, but little calculation to be done.
ENOUGH! that's just my $0.02