Aww... You lost me as soon as you started on the football stuff. With all this debacle I guess I'll fight through the sport references and read the article anyways :D
An opinion article states a conclusion that is based on some foundation of data that the author posits as facts. If those facts are... skewed... then the conclusion cannot help but be flawed. Positing that the G5 is not a 64-bit processor will *surely* be met with 'discussion', particulary since the common 'opinion' on the G5 is that it *is* a 64-bit processor.
About 64-bit and memory:
Also, 64-bit processing is not just about RAM... it's about the virtual address space. For example, being able to memory map a file that is 10G in size and step through it as if it were 'just memory' simplifies many tasks. This is useful even if your computer only has 512MiB of memory. In a 32-bit system, you have to do your own buffer manipulation, for example, to pull in parts of a large file and operate on each chunk. You have to deal with "oh, I need to go back and read the 2nd chunk now, now the 10th, now the 3rd, etc." With 64-bit OSs, you can just pretend that the file is a big array by mapping it into memory and traversing it just like any other chunk of memory in your program (as an array of some kind, for example). While it's nice to be able to have a machine with 8G memory all available to a process, being able to do things like map a huge file into your address space so you can manipulate it just like any other chunk of memory (and let the pagefile system and file system deal with caching which parts are in memory at the time when you don't have enough RAM to hold it all) are much nicer to deal with instead of having to write your own block processing algorithms.
The dynamic range that 64-bit integers off is nother area where 64-bit code helps. For example, in GIS, a 'normal' mode is to transform coordinates for an object based off of some offset that the grid is located in order to provide 'real world' coordinates. Integers may be used because they are very fast. 32-bit integers with an offset per grid is used because using only 32 bits doesn't map precisely enough (you cannot map the entire Earth down to a 'nice' resolution with only 32 bits). So, with 32-bits and an offset, you can map the grid down to a certain point then be more precise inside that grid. This is done so that all local operations to the grid use 32-bit mathmatic operations (which are fast).
While the more simple 64-bit mathmatical operations are handled well by 32-bit processors (add, subtract), the more complex ones are harder (multiply, divide, transcendentals/trigonometric) and slower. With a 64-bit coordinate system, you can map the entire planet down to a fairly nice resolution so you can go with a 'one-world coordinate system'. The 64-bit processors handle multiplication/division/transcendentals much faster than a 32-bit system doing 64-bit math. The "one-world' coordinate system then makes mapping software easier to write (no worrying about transposing one set of coordinates from one grid to a set of coordinates relative to an overlapping grid for some other system to use, for example).
Anyway, I'm sorry my other posts (and cross-link) caused you to feel as if you had to write this article. However, as you said yourself, when you present opinion and posit statements as fact on the Internet, *someone* will see it (at least you hope they do if you're writing for a publication) and if you have shaky or controversial points, you must expect to be called on them.
And... it I wouldn't call it 'elitism' disputing statements that someone else claims as fact. Questioning your posit of whether a G5 is 64-bit or not is not 'elitism', for example. There are ways to 'make it simple' without distorting the fact so much that it becomes false, or weakening it to the point where it means something else entirely, in the process.
Oh... addendum to the above... for the non-programmer/system person, I would agree that the main thing he will notice is that he can put more memory into his machine and for some things, the machine will run faster (and maybe be more stable for some applications that can use more simple algorithms due to 64-bit benefits). Most of 64-bit's 'goodness' is visible to programmers.
Originally Posted by fitten An opinion article states a conclusion that is based on some foundation of data that the author posits as facts. If those facts are... skewed... then the conclusion cannot help but be flawed. Positing that the G5 is not a 64-bit processor will *surely* be met with 'discussion', particulary since the common 'opinion' on the G5 is that it *is* a 64-bit processor.
[snipping irrelevant and off topic part that belongs in the other article]
Anyway, I'm sorry my other posts (and cross-link) caused you to feel as if you had to write this article. However, as you said yourself, when you present opinion and posit statements as fact on the Internet, *someone* will see it (at least you hope they do if you're writing for a publication) and if you have shaky or controversial points, you must expect to be called on them.
And... it I wouldn't call it 'elitism' disputing statements that someone else claims as fact. Questioning your posit of whether a G5 is 64-bit or not is not 'elitism', for example. There are ways to 'make it simple' without distorting the fact so much that it becomes false, or weakening it to the point where it means something else entirely, in the process.
Don't worry fitten, your posts hardly were the impetus for the article (though did partially fuel a discussion topic). My point is meant to be much more broad than this article. And though your first post on that article was exactly the type I'm talking about, my rant is more about a couple of the other "know-it-alls" in that discussion (unless you are the Hat guy over there).
It's elitism when someone tries to speak "down" at another person who's sharing his or her knowledge. Your G5 discussion is perfectly reasonable, but another person over there tried to prove he knew so much more than I did that he got several of his own facts wrong in the process. All for what? To say he knew more than I did about an article that was, if he were at the level of intellect he claimed, not targeted at him anyway? To argue nuances that were outside of the scope?
Just as there are ways to make some things simple, there are ways to dispute things without trying to be a know-it-all. That was my point in a nutshell, whether you agree or disagree. :) Whether it be against a writer, a fellow enthusiast, etc., a little courtesy and the realization that you, too, don't know everything goes a looooooooooooong way. :)
Interesting article and I feel where you are coming from. Iâve been knocking about for a while now (nearly 20 years in the hardware / software industry) and you find similar attitudes in real life as well as forums. I find as a general rule, people who really know what they are talking about are humble about their knowledge and happy to share that knowledge with anyone who shows genuine interest. This is in sharp contrast to what I call the three topic types, they invariably have learnt a bit about three basic subjects and like to try and impress with their limited knowledge at every available opportunity. Donât let it get to you, if you are flamed or insulted I can pretty much guarantee that with a little digging you will find that person really doesnât have a clue.
Originally Posted by richard parker Interesting article and I feel where you are coming from. Iâve been knocking about for a while now (nearly 20 years in the hardware / software industry) and you find similar attitudes in real life as well as forums. I find as a general rule, people who really know what they are talking about are humble about their knowledge and happy to share that knowledge with anyone who shows genuine interest. This is in sharp contrast to what I call the three topic types, they invariably have learnt a bit about three basic subjects and like to try and impress with their limited knowledge at every available opportunity. Donât let it get to you, if you are flamed or insulted I can pretty much guarantee that with a little digging you will find that person really doesnât have a clue.
QFT! That's been my experience also. A little knowledge is a dangerous thing as the saying goes.
Comments 26 to 32 of 32
ReplyAbout 64-bit and memory:
Also, 64-bit processing is not just about RAM... it's about the virtual address space. For example, being able to memory map a file that is 10G in size and step through it as if it were 'just memory' simplifies many tasks. This is useful even if your computer only has 512MiB of memory. In a 32-bit system, you have to do your own buffer manipulation, for example, to pull in parts of a large file and operate on each chunk. You have to deal with "oh, I need to go back and read the 2nd chunk now, now the 10th, now the 3rd, etc." With 64-bit OSs, you can just pretend that the file is a big array by mapping it into memory and traversing it just like any other chunk of memory in your program (as an array of some kind, for example). While it's nice to be able to have a machine with 8G memory all available to a process, being able to do things like map a huge file into your address space so you can manipulate it just like any other chunk of memory (and let the pagefile system and file system deal with caching which parts are in memory at the time when you don't have enough RAM to hold it all) are much nicer to deal with instead of having to write your own block processing algorithms.
The dynamic range that 64-bit integers off is nother area where 64-bit code helps. For example, in GIS, a 'normal' mode is to transform coordinates for an object based off of some offset that the grid is located in order to provide 'real world' coordinates. Integers may be used because they are very fast. 32-bit integers with an offset per grid is used because using only 32 bits doesn't map precisely enough (you cannot map the entire Earth down to a 'nice' resolution with only 32 bits). So, with 32-bits and an offset, you can map the grid down to a certain point then be more precise inside that grid. This is done so that all local operations to the grid use 32-bit mathmatic operations (which are fast).
While the more simple 64-bit mathmatical operations are handled well by 32-bit processors (add, subtract), the more complex ones are harder (multiply, divide, transcendentals/trigonometric) and slower. With a 64-bit coordinate system, you can map the entire planet down to a fairly nice resolution so you can go with a 'one-world coordinate system'. The 64-bit processors handle multiplication/division/transcendentals much faster than a 32-bit system doing 64-bit math. The "one-world' coordinate system then makes mapping software easier to write (no worrying about transposing one set of coordinates from one grid to a set of coordinates relative to an overlapping grid for some other system to use, for example).
Anyway, I'm sorry my other posts (and cross-link) caused you to feel as if you had to write this article. However, as you said yourself, when you present opinion and posit statements as fact on the Internet, *someone* will see it (at least you hope they do if you're writing for a publication) and if you have shaky or controversial points, you must expect to be called on them.
And... it I wouldn't call it 'elitism' disputing statements that someone else claims as fact. Questioning your posit of whether a G5 is 64-bit or not is not 'elitism', for example. There are ways to 'make it simple' without distorting the fact so much that it becomes false, or weakening it to the point where it means something else entirely, in the process.
Don't worry fitten, your posts hardly were the impetus for the article (though did partially fuel a discussion topic). My point is meant to be much more broad than this article. And though your first post on that article was exactly the type I'm talking about, my rant is more about a couple of the other "know-it-alls" in that discussion (unless you are the Hat guy over there).
It's elitism when someone tries to speak "down" at another person who's sharing his or her knowledge. Your G5 discussion is perfectly reasonable, but another person over there tried to prove he knew so much more than I did that he got several of his own facts wrong in the process. All for what? To say he knew more than I did about an article that was, if he were at the level of intellect he claimed, not targeted at him anyway? To argue nuances that were outside of the scope?
Just as there are ways to make some things simple, there are ways to dispute things without trying to be a know-it-all. That was my point in a nutshell, whether you agree or disagree. :) Whether it be against a writer, a fellow enthusiast, etc., a little courtesy and the realization that you, too, don't know everything goes a looooooooooooong way. :)
-
« Previous
-
1
-
2
-
Next »
Discuss in the forums