I ended up writing a python extension in C (painful when you've been programming in python).
Bridging between languages with totally different type systems is usually tedious. Try SWIG, or PyRex, or Boost Python.
Specifically, you can't compile python - and interpreted it's just too slow for anything computationally expensive.
Are you trolling or just ignorant? Sorry for being harsh but this statement is wrong, misleading, and overly broad all at the same time.
First, Python is NOT interpreted (yes I know some of the docs refer to the "Python Interpreter"...read on). It's compiled to byte code and then executed in a VM (just like Java and C# and etc). The closest thing to interpretation is using the interactive environment or maybe (if you really want to stretch) using 'exec' or the other functions that run strings as code. Even then it's still compiled and run in the VM. Just because there's no compile button or makefile doesn't mean compilation isn't happening.
As for being to slow because it's interpreted...re-read the above. If you think it's to slow because it's not native the same goes for Java and C#. Either way, what do you mean by "computationally expensive"? What application, specifically, have you found Python to be too slow. Are you talking about built-ins or libraries? A good portion of anything where speed would be a concern is already written in C. Highly optimized C at that. This is not to imply that you should write a 3D rendering engine in Python, but to point out that saying "it's too slow" without qualification is an old, tired, and disproven argument.
Also you have to load the interpreter, (or embed w/extra modules) it's slow and uses more memory than it should.
Again, what SPECIFICALLY are you talking about? Are you trying to embed or extend? If there are "extra modules" why don't you rebuild Python without them? How much memory was used in what situation and why do you think it was too much?
...even occasionally the need for typing...
Ugh...now I'm really starting to think this is just a troll...that notwithstanding:
Typing? Python is a strongly and dynamically typed language. If you don't know what that means than you shouldn't use "typing" as a reason for anything.
and assignment by reference.
Yet again, what are you talking about? Python always uses reference semantics.
No: There was reverse engineering done for the FAT file systems, but vastly less.
Yes: It's a much simpler file system.
More explanation:
FAT file systems just use a linked list of cluster locations to "tell" where on the disk the file is. The head of the list is in the files directory entry. That's about it. It has no indexes, no real performance features, and only one rudimentary (barely) integrity feature.
As such, FAT file systems have very few data structures involved with their inner workings. NTFS has, at the very least, dozens.
Also, since FAT file systems lack the performance features (specifically the indexes) there are no complicated algorthms to reverse engineer.
I'll buy that, but I got the impression that the OP was referring specifically to XP. XP was the first consumer branch of the NT line. It did include a point rev of the kernel, but it's not directly in the 'server' path. As such it's the first consumer edition that had a large internal change since 95.
Sure, if you'd abandoned the 9x consumer line and had been running 2k it wasn't that much of a change. But that wasn't the situation for most consumers.
Lot's of effort has been put into it. It's just an extraordinarily difficult thing to do. The amount of effort to reverse engineer/document the internal structures to support read access must have been huge.
The reason NTFS write access is so difficult to develop is not because because of the NTFS structures themselves, but because the algorithms that the file system driver uses are unknown.
For example, the details of balancing/re-balancing the b+trees that maintain the indexes. The file system driver has a bunch of criteria as to how many indexes should be in a node, what the maximum depth of the leaf nodes is, and etc.
That's just one of the algorithms that needs to be figured out in order to have safe write access. There are others (creating/maintaining the data runs, managing resident and non-resident attributes, etc) Figuring these out, with all their special cases and boundary conditions, is difficult. You can either try to make a bazillion tests and hope you catch all the weird corner cases (which is hard, slow, and you never know when you're done) or you can completely reverse engineer the file system driver (also not too easy).
The consequences of screwing it up are also hard to fully figure out. At best, maybe you just get sucky performance, at worst you completely destroy directories and files.
It's a tough job just to implement a read, so it makes sense that writes haven't come as far.
Those Quebec people make some damn fine beer. I think I've had most of their offerings (including a couple of hard to get ones) and I agree, Le Fin du Monde is one of my favorite beers.
You do realise that that is the argument used for centuries to keep blacks and women from voting, right?
Yes I do. Keeping people out of the political process for unrelated and asinine reasons was a travesty. Unfortunately, the fight (and victory) against these injustices left us with the idea that ANY form of political descrimination was reprehensible. I don't find this to be the case. Most people take their citizenship for granted and don't maintain a level of familiarity and understanding regarding current social/political issues such that they can reasonably contribute.
I agree with the low opinion of politicians, but that doesn't make a pure democracy any better.
In my opinion, both of our complaints regarding the current state of affairs can be traced back to the decay/loss of an informed electorate. Only an informed and educated electorate can sustain the benefits attributed to democracy.
Unfortunately, informed and educated electorates are hard to get in the first place and even harder to maintain. Until such an electorate is the norm, a representative system is the better choice.
Granted, our current representatives ain't exactly Plato's 'Philosopher Kings'.
-snip- [I don't like the dude either, but that's not my point...]
Why don't I support a pure democracy? Jefferson said it best:
'A democracy is nothing more than mob rule, where fifty-one percent of the people may take away the rights of the other forty-nine.'
I do not believe that citizenship makes one a good policy maker. All people should be equal under the law (another current problem). Beyond that, all bets are off. I do not wish to entrust my, or anyone elses, civil rights to a simple majority.
Which is exactly why we (the United States) had (initially) a democratic republic as opposed to a pure democracy. Some/most U.S. citizens are:
stupid/ignorant/evil/jerk-offs/small-minded/"can 't think for themselves"/un-informed.
Just because they are citizens does not mean that they should be allowed to make policy. Viz, just because they happen to be human doesn't mean their opinion is useful, important, or valuable.
"From the advent of the player piano, every new means of reproducing sound has struck a dissonant chord with musical copyright owners, often resulting in federal litigation."
I agree, but I think that this is compatible with his (Paul Graham) statements. Great programmers will recognize situations where one language will be better than another. 'Regular' programmers will just bash square pegs through round holes...or give up on the problem entirely.
A myth? Hardly. The linkage jamming argument is often used in a misleading fashion. Yes, it's true that key placement was selected to avoid jamming...jamming that happened more often at higher speeds. Even if you don't count that as being 'to slow down typists' (which is arguable) it's still the net effect.
The study you refer to was considered suspicious by other researchers. When attempts were made to look at the raw data it turned out it had been destroyed (details). A single study with no access to the raw data isn't very convincing.
The usual study people mention as proof of qwerty's superiority was written by economists. Also not terribly convincing. Here's a refutation.
Either way, it's clear that the main design decision with qwerty was to in response to an engineering problem.
The best way to find out is to try it. I used to quote the same studies and had the same opinions until I tried it myself. It took about a weekend to get started and three weeks to exceed my previous level of proficiency.
[The Scene: a writer at their desk with deadline looming:]
Must have idea...
Must get editorial in on time...
Politics are done...
Not much happening in the literature scene...
Chomksy stuff is too complicated...
Must be controversial...but not too controversial to the prime demographic...
Whoa...I've got it!!!
[writer bites tongue and begins scribbling onto a ruled notebook, we see the title:]
"The Software Industry is Dead!"
[writer scribbles madly for 90 minutes, has a lot of rough red wine during the scribbling, and then falls asleep on the draft (and dreams of Hemingway).
There are restrictions on both -- in the US, freedom of speech is not an absolute. Under the GNU GPL, your freedom to share and modify comes with a proviso that you cannot deny recipients of distributed derivatives or verbatim copies the freedoms to share and modify the program.
I really like this explanation. I think it accurately and succinctly describes the situation. Unfortunately, many vocal proponents of Free Software put the emphasis on the recipients freedoms; statements of "not as in beer" notwithstanding. As you describe above, the freedom comes from not having one party restrict another partys rights down the line.
Not at all. The FSF uses driving a car to help understand why restricting some freedoms are necessary to preserve others; I'll attempt to paraphrase it briefly: we cannot have all possible freedoms because some conflict. So we make choices and give up some freedoms to keep other freedoms. For instance, we are not allowed to drive anywhere we want at any speed we want. We are not allowed to drive on the sidewalks and we are not allowed to disobey the speed limit. Our freedom to do these things is curtailed because other freedoms are deemed more valuable -- the freedom to walk down the street in safety. The GNU project is about spreading software freedom to more people, so this requires a copyright license which doesn't allow anyone to strip away the freedoms of free software. Hence the GNU GPL (the license under which a lot of the GNU project's programs are distributed) has a strong copyleft.
I agree with your paraphrase, again I think it well describes the issue. However, I personally disagree with the underlying philosophy. That's not to say that I don't like the GPL or any other comment as to the merits of any of the various licenses mentioned; it's that I've never been a big fan of the Rousseau-style ideas about freedom. Regardless of my thoughts on "social contract" or any other means of getting along, I prefer to describe freedoms proactively: "you can do this". Not in the "freedom from" fashion.
Granted, those are just personal preferences. But I think that using the word free in a more "rarified" sense than it normally is often results in very facile analogies that are really just sound bites. This is unfortunate. I think the concepts, as you outline them above, are important and worthy of discussion. But it is my opinion that the decision to use the word "free" was rather misleading. It sounds good, but the sense it was used in requires the acceptance of a certain philosophical position that is not immediately apparent. The FSF and some proponents may describe the situation in a forthright and accurate manner but what seems to happen more often is the stating of simplistic arguments along the lines of "freedom being good like speech but different than beer".
Re:How many licenses can fit on the head of a pin?
on
PHP Not Moving To The GPL
·
· Score: 5, Insightful
The GPL is free as in free speech, meaning "freedom".
What? How does "as in speech" make your point here? I don't even see how it's related. Yes, I know this is a GPL mantra...I even understand it, but in your example I don't see the connection.
With freedom, as we all know, comes responsibility.
Where do you get that? I don't follow the logical leap here. I know it's a common thing to say but how does it apply to your argument?
The PHP licence is only "more free" than the GPL if you use a different meaning of "free" than that which the GPL is based on.
This sounds like double-talk to me. I understand the rationale behind the GPL. I may go so far as to say that I agree with its intent in some cases, but it's a bit disingenuous to try to convince someone that a more restrictive license is somehow more "free".
Making statements like this just confuses people even further, so please don't do it.
It's confusing because of the issues mentioned in my above comments.
Slashdot Mirror
I ended up writing a python extension in C (painful when you've been programming in python).
...even occasionally the need for typing...
Bridging between languages with totally different type systems is usually tedious. Try SWIG, or PyRex, or Boost Python.
Specifically, you can't compile python - and interpreted it's just too slow for anything computationally expensive.
Are you trolling or just ignorant? Sorry for being harsh but this statement is wrong, misleading, and overly broad all at the same time.
First, Python is NOT interpreted (yes I know some of the docs refer to the "Python Interpreter"...read on). It's compiled to byte code and then executed in a VM (just like Java and C# and etc). The closest thing to interpretation is using the interactive environment or maybe (if you really want to stretch) using 'exec' or the other functions that run strings as code. Even then it's still compiled and run in the VM. Just because there's no compile button or makefile doesn't mean compilation isn't happening.
As for being to slow because it's interpreted...re-read the above. If you think it's to slow because it's not native the same goes for Java and C#. Either way, what do you mean by "computationally expensive"? What application, specifically, have you found Python to be too slow. Are you talking about built-ins or libraries? A good portion of anything where speed would be a concern is already written in C. Highly optimized C at that. This is not to imply that you should write a 3D rendering engine in Python, but to point out that saying "it's too slow" without qualification is an old, tired, and disproven argument.
Also you have to load the interpreter, (or embed w/extra modules) it's slow and uses more memory than it should.
Again, what SPECIFICALLY are you talking about? Are you trying to embed or extend? If there are "extra modules" why don't you rebuild Python without them? How much memory was used in what situation and why do you think it was too much?
Ugh...now I'm really starting to think this is just a troll...that notwithstanding:
Typing? Python is a strongly and dynamically typed language. If you don't know what that means than you shouldn't use "typing" as a reason for anything.
and assignment by reference.
Yet again, what are you talking about? Python always uses reference semantics.
Short answers:
No: There was reverse engineering done for the FAT file systems, but vastly less.
Yes: It's a much simpler file system.
More explanation:
FAT file systems just use a linked list of cluster locations to "tell" where on the disk the file is. The head of the list is in the files directory entry. That's about it. It has no indexes, no real performance features, and only one rudimentary (barely) integrity feature.
As such, FAT file systems have very few data structures involved with their inner workings. NTFS has, at the very least, dozens.
Also, since FAT file systems lack the performance features (specifically the indexes) there are no complicated algorthms to reverse engineer.
I'll buy that, but I got the impression that the OP was referring specifically to XP. XP was the first consumer branch of the NT line. It did include a point rev of the kernel, but it's not directly in the 'server' path. As such it's the first consumer edition that had a large internal change since 95.
Sure, if you'd abandoned the 9x consumer line and had been running 2k it wasn't that much of a change. But that wasn't the situation for most consumers.
I wonder why not more effort has been put into it
Lot's of effort has been put into it. It's just an extraordinarily difficult thing to do. The amount of effort to reverse engineer/document the internal structures to support read access must have been huge.
The reason NTFS write access is so difficult to develop is not because because of the NTFS structures themselves, but because the algorithms that the file system driver uses are unknown.
For example, the details of balancing/re-balancing the b+trees that maintain the indexes. The file system driver has a bunch of criteria as to how many indexes should be in a node, what the maximum depth of the leaf nodes is, and etc.
That's just one of the algorithms that needs to be figured out in order to have safe write access. There are others (creating/maintaining the data runs, managing resident and non-resident attributes, etc) Figuring these out, with all their special cases and boundary conditions, is difficult. You can either try to make a bazillion tests and hope you catch all the weird corner cases (which is hard, slow, and you never know when you're done) or you can completely reverse engineer the file system driver (also not too easy).
The consequences of screwing it up are also hard to fully figure out. At best, maybe you just get sucky performance, at worst you completely destroy directories and files.
It's a tough job just to implement a read, so it makes sense that writes haven't come as far.
A new service pack, just like XP was
Incorrect. XP was the first release since 95 that had major internal changes. XP was the first consumer edition based on the NT kernel.
Damn it.
Yup...it's 'La'
I really did check first, I just checked a source that wasn't any better at French than I am...
Thanks for the correction.
...but what they are doing a fine job of is cloning Belgian beer...
;)
This is true...for some reason I thought they were quite up front about the fact that the styles are Belgian...maybe I was thinking of Ommegang.
Then again, anyone who knows Belgian-style beer can already tell
Those Quebec people make some damn fine beer. I think I've had most of their offerings (including a couple of hard to get ones) and I agree, Le Fin du Monde is one of my favorite beers.
You do realise that that is the argument used for centuries to keep blacks and women from voting, right?
Yes I do. Keeping people out of the political process for unrelated and asinine reasons was a travesty. Unfortunately, the fight (and victory) against these injustices left us with the idea that ANY form of political descrimination was reprehensible. I don't find this to be the case. Most people take their citizenship for granted and don't maintain a level of familiarity and understanding regarding current social/political issues such that they can reasonably contribute.
I agree with the low opinion of politicians, but that doesn't make a pure democracy any better.
In my opinion, both of our complaints regarding the current state of affairs can be traced back to the decay/loss of an informed electorate. Only an informed and educated electorate can sustain the benefits attributed to democracy.
Unfortunately, informed and educated electorates are hard to get in the first place and even harder to maintain. Until such an electorate is the norm, a representative system is the better choice.
Granted, our current representatives ain't exactly Plato's 'Philosopher Kings'.
-snip-
[I don't like the dude either, but that's not my point...]
Why don't I support a pure democracy? Jefferson said it best:
'A democracy is nothing more than mob rule, where fifty-one percent of the people may take away the rights of the other forty-nine.'
I do not believe that citizenship makes one a good policy maker. All people should be equal under the law (another current problem). Beyond that, all bets are off. I do not wish to entrust my, or anyone elses, civil rights to a simple majority.
Which is exactly why we (the United States) had (initially) a democratic republic as opposed to a pure democracy. Some/most U.S. citizens are:
n 't think for themselves"/un-informed.
stupid/ignorant/evil/jerk-offs/small-minded/"ca
Just because they are citizens does not mean that they should be allowed to make policy. Viz, just because they happen to be human doesn't mean their opinion is useful, important, or valuable.
Amiga
"From the advent of the player piano, every new means of reproducing sound has struck a dissonant chord with musical copyright owners, often resulting in federal litigation."
Get down with the catchy metaphor...I like it.
Is this normal style in opinions or is this rare?
...I was reading this thing we had called slashdot, as was the style of the time, when we heard that the GPL was gonna get tested in court...
I agree, but I think that this is compatible with his (Paul Graham) statements. Great programmers will recognize situations where one language will be better than another. 'Regular' programmers will just bash square pegs through round holes...or give up on the problem entirely.
They're generally smart enough to figure out that setting it on a laptop isn't the main idea.
Have you seen the touchstream by fingerworks? I got one of these right after I first saw it for that same reason. Expensive, but cool.
A myth? Hardly. The linkage jamming argument is often used in a misleading fashion. Yes, it's true that key placement was selected to avoid jamming...jamming that happened more often at higher speeds. Even if you don't count that as being 'to slow down typists' (which is arguable) it's still the net effect.
The study you refer to was considered suspicious by other researchers. When attempts were made to look at the raw data it turned out it had been destroyed (details). A single study with no access to the raw data isn't very convincing.
The usual study people mention as proof of qwerty's superiority was written by economists. Also not terribly convincing. Here's a refutation.
Either way, it's clear that the main design decision with qwerty was to in response to an engineering problem.
The best way to find out is to try it. I used to quote the same studies and had the same opinions until I tried it myself. It took about a weekend to get started and three weeks to exceed my previous level of proficiency.
Same thing (pancreatic cancer...not necessarily the rare kind) got him. Though I doubt he's all that peaceful; all things considered.
N.B. No one should get this hideous mallady.
Is this actual demolition/weapons/forensic lingo or is this just supposed to sound 'informed'?
[The Scene: a writer at their desk with deadline looming:]
Must have idea...
Must get editorial in on time...
Politics are done...
Not much happening in the literature scene...
Chomksy stuff is too complicated...
Must be controversial...but not too controversial to the prime demographic...
Whoa...I've got it!!!
[writer bites tongue and begins scribbling onto a ruled notebook, we see the title:]
"The Software Industry is Dead!"
[writer scribbles madly for 90 minutes, has a lot of rough red wine during the scribbling, and then falls asleep on the draft (and dreams of Hemingway).
There are restrictions on both -- in the US, freedom of speech is not an absolute. Under the GNU GPL, your freedom to share and modify comes with a proviso that you cannot deny recipients of distributed derivatives or verbatim copies the freedoms to share and modify the program.
I really like this explanation. I think it accurately and succinctly describes the situation. Unfortunately, many vocal proponents of Free Software put the emphasis on the recipients freedoms; statements of "not as in beer" notwithstanding. As you describe above, the freedom comes from not having one party restrict another partys rights down the line.
Not at all. The FSF uses driving a car to help understand why restricting some freedoms are necessary to preserve others; I'll attempt to paraphrase it briefly: we cannot have all possible freedoms because some conflict. So we make choices and give up some freedoms to keep other freedoms. For instance, we are not allowed to drive anywhere we want at any speed we want. We are not allowed to drive on the sidewalks and we are not allowed to disobey the speed limit. Our freedom to do these things is curtailed because other freedoms are deemed more valuable -- the freedom to walk down the street in safety. The GNU project is about spreading software freedom to more people, so this requires a copyright license which doesn't allow anyone to strip away the freedoms of free software. Hence the GNU GPL (the license under which a lot of the GNU project's programs are distributed) has a strong copyleft.
I agree with your paraphrase, again I think it well describes the issue. However, I personally disagree with the underlying philosophy. That's not to say that I don't like the GPL or any other comment as to the merits of any of the various licenses mentioned; it's that I've never been a big fan of the Rousseau-style ideas about freedom. Regardless of my thoughts on "social contract" or any other means of getting along, I prefer to describe freedoms proactively: "you can do this". Not in the "freedom from" fashion.
Granted, those are just personal preferences. But I think that using the word free in a more "rarified" sense than it normally is often results in very facile analogies that are really just sound bites. This is unfortunate. I think the concepts, as you outline them above, are important and worthy of discussion. But it is my opinion that the decision to use the word "free" was rather misleading. It sounds good, but the sense it was used in requires the acceptance of a certain philosophical position that is not immediately apparent. The FSF and some proponents may describe the situation in a forthright and accurate manner but what seems to happen more often is the stating of simplistic arguments along the lines of "freedom being good like speech but different than beer".
Thats absurd, freedoms do entail responsibility.
Why?
The GPL is free as in free speech, meaning "freedom".
What? How does "as in speech" make your point here? I don't even see how it's related. Yes, I know this is a GPL mantra...I even understand it, but in your example I don't see the connection.
With freedom, as we all know, comes responsibility.
Where do you get that? I don't follow the logical leap here. I know it's a common thing to say but how does it apply to your argument?
The PHP licence is only "more free" than the GPL if you use a different meaning of "free" than that which the GPL is based on.
This sounds like double-talk to me. I understand the rationale behind the GPL. I may go so far as to say that I agree with its intent in some cases, but it's a bit disingenuous to try to convince someone that a more restrictive license is somehow more "free".
Making statements like this just confuses people even further, so please don't do it.
It's confusing because of the issues mentioned in my above comments.