When that code needs to be modified to run on a supercomputer instead of a workstation, it is usually converted to pure C or Fortran.
Why is this done? If your Python program is spending almost all of its time inside a C module, why is it worth converting the whole thing to C? Not a flame, I just don't understand. Is it because you run the whole computation over and over in a loop, and the Python overhead starts to be noticeable?
P.S. You are a mathematician... have you seen Sage? It's basically a whole bunch of math libraries, glued together with Python.
I wonder if Fortran may eventually be replaced by Python.
A few years ago, when I was first getting into Python, I read an article where a guy from a science research lab talked about his lab's transition from Fortran to Python. Python has some nifty heavy-duty math modules, written in C; and everyone at the lab who tried out the Python stuff strongly preferred it to Fortran.
Since C code is doing all the heavy lifting, it's nice and fast. Since Python is interactive, scientists can use it as a really-powerful desk calculator. And since Python has a clean and friendly syntax, it's easier to write and debug Python programs than Fortran.
I really wish I had saved a copy of that article, or at least its URL. I've tried Google searching for it, and I find many hits on using Python in labs but I haven't found the article.
Sure, that works. It's still a kludgey workaround that results in you doing part of the compiler's job.
Yet, it's the best I have come up with do to this in Python. I already stipulated that other languages (specifically, Lisp) are easier with respect to this.
Do you have a better way to do it in Python? Or a way to change Python to make this easier... that might actually get accepted into Python? If not, you're just complaining. Which you are of course free to do!
Likewise, if a function is only needed in one place (say, as a callback for a UI event), it's often both more convenient and readable to just stick it inline.
You sound like you think it would be sort of nice to have, which is about where I am on this issue. But some people seem to really care about this.
Python isn't perfect, but in my opinion it's about 95% of the way there and I'm willing to live with the warts (such as the closures hack I documented). I haven't found anything I like better than Python for high-level coding.
If you need a function that does two things, you can't use lambda anymore. This is not a great hardship as Python allows you to declare inner-scoped functions and you can use that instead,
I'd like to call attention to this point, since everyone but you seems to have missed it.
In Python, lambda expressions are so limited that they are really only useful for trivial purposes. People have asked Guido to improve lambda, but he hasn't. My understanding is that Python's parser somehow makes it hard to parse a multi-line lambda.
The solution is very easy, however.
The purpose of lambda is to create an anonymous function object, often with a closure. Well, in Python, it's trivial to create a function object by declaring a function; you can declare a function anywhere, including in a nested scope, and you can make closures. The only thing you can't do is declare a function with no name. So all you really need is a coding convention, such as: every function I wish was a lambda function will be called "anon_fn".
Here's an example. We want to make a function that returns anonymous function objects that are adders: we will pass in a value x, and we want our adder function objects to remember that x, add it to some number n, and return the result. Once we have make_adder() we can make as many anonymous adder function objects as we wish.
def make_adder(x):
def anon_fn(n):
return x + n
return anon_fn
The above example is trivial enough you could do it with a lambda:
def make_adder(x):
return lambda n: x + n
But I prefer the Python version. In fact, I wouldn't actually use "anon_fn" as the name, I'd probably call the function "adder" and say "return adder".
def make_adder(x):
def adder(n):
return x + n
return adder
make_adder() creates a function object. For a little while, inside make_adder(), this function object is bound to the name "adder", but that binding vanishes when make_adder() returns. The function object is returned, and now it's anonymous.
So, to me the interesting question is not "when will Python get multi-line lambda?" To me, the question is "Why is it so important to be able to make a function that never was bound to a name for even a moment?" This isn't a flame, I'm actually wondering why this seems like such a big deal to so many people.
P.S. As long as I'm beating the lambda horse, I might as well beat the closures horse as well.
People argue that Python doesn't have "true" closures. I think it does, but the way it works may not be obvious.
In Python, variables are really just names that have objects bound to them. If you think of the closure as the name, then closures are broken in Python. If you think of the closure as the object, not so broken.
Here we want to make a function that returns anonymous function objects that are counters. Each counter function object should have an initial value, enclose that value, and when given a number n, add n to the current enclosed value and return the updated value.
# this version seems reasonable but does not work def make_counter(initial_value):
x = initial_value
def counter(n):
x += n
return x
return counter
What's wrong here? The attempted closure is trying to use a name, "x", where Python needs an object. "x += n" really means "rebind the name 'x' with its previous value plus the value of 'n'". But when you enter counter(), the name x is not bound. Python will enclose the object bound to a name, but doesn't really keep the name binding part.
Over time, all the cool features from proprietary UNIX versions are getting ported to Linux, either directly or by being re-implemented. As Linux becomes more and more acceptable as a replacement, expect to see proprietary UNIX versions start to go away.
If IBM hires a person to work on Linux, that work helps IBM across pretty much their whole product line. If IBM hires a person to work on AIX, that work has much less value now, and will have even less and less value over time as Linux gathers up more of the market. Also, as Linux keeps getting better, it would take more and more work to add similar features to AIX, to try to keep up. Eventually, IBM is going to stop paying for work on AIX at all; they will end-of-life AIX, and just sell Linux.
I don't know for sure about SMIT but Linux does have LVM and various tools to manage it. AIX gurus, how ready is Linux to replace AIX now?
And, are desktop POWER machines going to be available with Linux?
It's crazy cool. He shows you how to make your own working spectroscope with a box, a CD, two razor blades, and some tape!
The guy who put up that site has written some actual paper books, so you could give one or more of those. Or, just order some magnets and diffraction gratings and such for building the gadgets, from the catalog:
Antivirus is one of those things that(at least until actual heuristic scanning that seriously works comes out) leans heavily on having a whole bunch of security guys and worker drones hammering out signature updates all day every day. That isn't something that falls under "The Open Source is strong with this one".
Hmmm, not sure I agree. I have always thought that the open source community could do a great job with antivirus.
The key is to get a large community of people who, when they discover a new virus, contribute their knowledge back to the open source project. And I think this is actually working with ClamAV.
I know that I have submitted my share of viruses... when I get an email offering me a cool new screen saver, and the file is called "screensave.scr.exe", I scan it with ClamAV. If ClamAV doesn't spot anything wrong, I'll submit that file to the ClamAV project.
Usually I submit the file at VirusTotal first, and attach the report to my submission.
ClamAV gets signatures very quickly for new viruses as they appear. The whole signature-based game is a continual game of catchup, though. I agree that heuristic-based scanning would be preferable, but that seems like a hard problem.
But have you ever considered that the culture might be less violent without a gun culture?
Statistics show that most civilian gun owners are law-abiding and non-violent. I quote Wikipedia:
Permit holders are a remarkably law-abiding subclass of the population. Florida, which has issued over 1,346,000 permits in twenty years, has revoked only 165 for a "crime after licensure involving a firearm," and fewer than 4,200 permits for any reason.
Most of the violent criminals are already convicted felons, and thus legally barred from owning a gun. I'd say they are part of the criminal culture, not the gun culture.
Have you actually researched the gun culture of the USA? I invite you to read The Gun Culture and Its Enemies. Amazon is offering a used copy for $0.44 plus shipping.
I've seen you posting on Slashdot, and I usually respect what you write. However, I completely disagree with your conclusions here, and I urge you to think about this some more.
No gun ban has ever succeeded in keeping guns out of the hands of violent criminals. They want guns, so they get them.
Some people will tell you that we could keep guns out of the hands of bad guys if we had a consistent, national, Draconian ban. A War on Guns, if you will. Think for a moment about the War on Drugs. Have we been able to keep crack cocaine out of the hands of users? If not, is it because we don't have a nation-wide Draconian ban on crack cocaine? Now consider that the crack users need to buy more every week. They can get drugs every week, in the face of a nation-wide ban. A violent thug only needs to get a gun once, and then he will carry it for months. (Basically, until he throws it away because cops are after him or something... then he gets another one.)
And, consider how easy it is with a machine shop to make a functioning firearm. Even if all gun makers were driven out of business, and all firearms circulating in society were found and destroyed, and all gun smugglers were caught at the border... all of which I consider impossible, by the way... even then, the bad guys would start making their own guns.
So, it is an axiom with me that no amount of law can keep dangerous weapons out of the hands of the bad guys. What then?
Then, all you can do is disarm the law-abiding. Does that really help?
Now, you mentioned Europe. I urge you to read the book The Samurai, the Mountie, and the Cowboy. I can't summarize a whole book in a few words, but the important takeaway is that violence is much more a cultural thing than a product of laws. England had a low violent crime rate. Then they banned guns. Then they had a low violent crime rate. People will tell you that England banned guns and has a lower crime rate than the USA; that's true, but can you really say it was the ban that caused the lower crime rate? (And in recent years, their violent crime rate has climbed...)
So, I don't believe that you can successfully disarm the bad guys. And I don't believe that you can lower the violent crime rate by passing gun control laws. And history and the law agree that the cops cannot protect everyone; all attempts to sue the police for failing to prevent any crime will fail, even if the police were horribly negligent. Given all this, I strongly oppose any and all efforts to ban firearms in civilian hands.
Also, bad neighborhoods in bad parts of town are where innocent people need firearms the most! I live in a boring, safe suburb, and I don't really need to own a gun. The truly poor people who live in horrible places are in much worse danger than I am. This is one reason I'm opposed to mandatory six-week training classes, expensive licenses, and bans on so-called "Saturday Night Specials". These things do nothing to stop the bad guys, and might keep an innocent citizen from getting a gun in a time of true need. I've read a few stories that curdled my blood.
There are a whole bunch of "history" tricks, to recall old commands without using the mouse.
When I started college, I studied the shell's man page until I knew them all. Some are so obscure I have forgotten them.
Generally, these involve an '!' character in some way.
Here are a few I use:
!! # run again the last command that was run !9 # run again the command with history number 9 !v # run again the last command that started with a 'v' !vi # run the last command that started with "vi" !?foo? # run the last command that had the string "foo" anywhere in it
diff oldfile newfile mv !$ !^ # same as "mv newfile oldfile" # !$ is last arg of previous command, !^ is first arg
ls foo bar baz rm -f !!* # same as "rm -f foo bar baz" # !!* repeats all arguments from previous command
There are actually some baroque tricks that recall a previous command and perform a search-and-replace on it, but for anything that complicated I just recall the line and edit it. The baroque tricks would have been pretty darn cool back in the paper teletype days, though.
By the way, the Bash shell can be configured to edit command lines using vi or Emacs commands. I described how to do it in an article I wrote for Linux Journal magazine. It's the last section, "vi or Emacs Mode in the Shell".
Oh, not exactly a history trick, but here's something I use all the time: ls -1 >/tmp/files vi/tmp/files # edit list to include just the files I want rm `cat/tmp/files` # `cmd` inserts the standard output from cmd into the command line as if you typed it
ls -1 >/tmp/files vi/tmp/files # edit list to include just the files I want # now run this command::%s+.*+mv &/some/directory/path/&+ # save file and quit vi source/tmp/files
This moves the chosen files to "/some/directory/path". The breakdown of the vi command is as so: : # invoke "ex mode" for search and replace command % # run the following command on every line of the file s # do a search and replace + # use a '+' for the command delimiter, so I won't have to backslash escape '/' chars in the path
.* # all characters on the line + # end the match pattern, begin replace pattern # & refers to the match pattern, thus all chars on the line mv &/some/directory/path/& # replace "foo" with "mv foo/some/directory/path/foo"
Takes less time to do it than to explain it!
The above is perhaps overkill if all the files are going to the same place. It's great if you want to send some files one place, some to another, because you can just edit the destinations until it looks right.
That "nothing has been proven" comment is pretty clearly tongue-in-cheek, just like the "take over the Internet" part.
The video encoding field is crowded with patents, so it's probably impossible to do something like Theora without needing a patent license. But Theora is based on some patented technology (VP3) whose patents have been donated for free use, irrevocably forever. And Theora is free, open source software with a BSD license. If you use the Ogg container format, Theora video, and Vorbis audio, you have a completely free media format.
So, you can use Theora for any purpose, without needing to pay royalties, without needing to get permission. That's why it's so funny that Nokia claimed Theora is "proprietary"... I do not think that word means what they think it does.
This is an actual car, but the Bond version remains sheer fantasy. The Rinspeed's passenger compartment is not pressurized; it's designed to let the water in. According to the above cited text:
With an enclosed volume of just two cubic meters of air the vehicle weight would have to increase by two tons (!) to counteract the unwanted buoyancy, giving the "sQuba" the land mobility of a turtle.
The James Bond movie car drove fast on land, and shot wet cement onto the windscreen of a pursuing car, before driving into the sea and then firing a missile to shoot down a helicopter. This is cooler, though, because it actually exists.
I recommend you just develop a web-based solution using Django. (Or any other web platform of your choice, but Django is the one I have used and I love it. Django makes it easy to do whatever I have needed to do, and I have done some unusual stuff with it.)
Then, you can use any phone with a decent web browser, plus you could use a PDA, a netbook, or anything else. You likely will live in the house for many years, and technology certainly will evolve... a simple web solution is pretty future-proof.
Google is your friend. There are so many different groups doing Kinetic Sculpture Races... do some searches, find their various web pages. If you live near one of the big races, you might be able to meet people in real life.
But what the heck, I did a few Google searches for you. Here's a few links:
Several of those web sites have mailing lists (maybe just for news of the specific race, maybe for general kinetic sculpture discussion, who knows). The first link has a mailto: link from the author ("Elliott Naess") and you could try sending him email to ask his advice.
That http://kinetickingdom.com/ site is a major resource. It's from the oldest kinetic sculpture racing group, in Ferndale, CA.
The guys who do this build wacky and weird human-powered contraptions. These are supposed to go on land, water, sand, and mud, and be only human-powered. (Some vehicles fail in one or more of these categories; but they can avoid being disqualified by providing sufficient bribes to the Kinetic Kops. In plain sight of all onlookers, of course.) Vehicles that can do all of the above, without any "pilots" leaving the vehicle to adjust things, get the "ACE Award" for good engineering.
One of my favorite kinetic sculpture vehicles is a behemoth that carries four people, each of whom provides power to one wheel, and one of whom has the steering wheel and brakes. I have also seen a vehicle that carried eight people, all powering a common drive train.
Anyway, these races have been happening for decades, and you can find the discussion lists where the KSR community discusses where to get parts, how to make things strong and reliable, etc.
I want it too. When I used to run Windows 98 with Norton Utilities, I had this ("Norton Protected Recycle Bin") and it was slick. See my 6-year-old comment:
To do this properly, it should be in the file system. (The undelete tool and the expire-old-data tool can and should be in user space, but the data protection features should be in the file system.)
I think the main thing I see is that the old argument, "Scripting languages are far too slow!" has finally been put to rest.
Well, that was always an oversimplification anyway. Too slow for what, specifically?
Even today, no one would seriously think about writing a video encoder entirely in a dynamic interpreted language. That's a very compute-intensive application and you can't afford the overhead. But how many of us write video encoders? There are many tasks for which the overhead of a dynamic interpreted language is no big deal.
Computers are really fast these days, so you can afford some overhead. If your trivial program runs in 0.6 seconds instead of 0.01 seconds, you may not care about the difference. And if you can write your program in 1/10 the time, you may come out way ahead. (And if the program is a one-off, that only needs to be run a few times, all you really care about is how much of your time it took to write the thing and get it correct.)
Instead of spending $300 on Bose headphones with active noise cancellation, consider spending $80 to $100 on some headphones that simply block background noise very well.
As I type this, I am listening to music on my Sennheiser HD-280 Pro headphones. According to Sennheiser they provide 30 dB of attenuation for background noise. They simply seal around your ears and passively block noise.
I was talking to a world-class expert on audio stuff one time, and if I recall correctly he was concerned about the noise canceling headphones possibly having a negative effect on your hearing. The active noise cancellation works by measuring the noise, and generating more noise with the phase opposite; if all goes well, the noise and the anti-noise cancel. But the active cancellation is pumping more energy out right next to your ear.
I'm not 100% certain I am remembering correctly what he said. I mean, if you wear the active cancellation headphones and the noise sounds quieter, doesn't that mean that less energy is reaching your ears? I'll have to ask him when I see him again. But that will be too late for this Slashdot discussion.
But you can spend less money to get a simpler device that passively blocks noise, so it seems like the way to go in any event.
By the way, I am quite satisfied with the audio quality of my HD 280 Pro headphones. They are the best headphones I have owned so far.
Over time, computers have become more and more powerful. And people always find uses for the additional power. If anyone tells you that we "don't need more power", ignore them.
But the netbooks show that there is a solid niche for people doing basic tasks wherever they want to go. I love going to small local espresso shops, and enjoying some sort of coffee drink while reading Slashdot. Even my OLPC XO is adequate for this, let alone an Atom-powered netbook. If anyone tells you that netbooks "aren't powerful enough", ignore them.
In fact, netbooks are powerful and have huge screens compared to PDAs and phones, and web sites shouldn't ignore PDAs and phones either. All the major web sites need a three screen strategy: their sites should work on TVs for living-room users, the site should work on PCs, and their sites should work on phones.
P.S. I was amused by the example from the summary. Video editing? Oh noes, my netbook is underpowered for video editing. How many people thought otherwise?
I was shocked by how excellent the "rough Google translation" was. Unless they had a human clean up the translation a bit, that is amazingly good English prose for a machine translator to emit. (I can't speak for how accurate it is, but it seems plausible enough.)
English is a mess, with lots of irregular usages. How about Norwegian -- is it particularly easy or particularly hard to translate?
"Do not be afraid to throw it away and start again."
I pretty much agree with most of your points. However, if you have done the left-corner design properly, you should have important building blocks ready to re-use when you do scrap some work and start again.
The audio DSP engine originally had to write the same exact audio format that it originally read; if you fed it 2-channel (stereo) data, you had to get 2-channel data out. When I rewrote it to add the mixer feature, I had to add quite a bit of code. But my code for reading and writing wave audio was perfectly re-usable, as was the actual DSP function code.
25 years ago I read a book called Software Tools in Pascal. This had a huge, beneficial impact on me.
The most important single thing I learned from that book is something they called "left-corner design". It goes like this:
Find some small part of your project, preferably something at the lowest design layer. Then code it up and implement it. Make it so brain-dead simple that you can spot all the bugs, and fix them. Now, consider some simple way to make it do something more, something else needed for the final product. Then make it do that, and fix it until the new feature is also working perfectly. Iterate.
As a real-life example, I once made an advanced audio DSP (digital signal processing) engine. It started out as a program that could open a wave audio file, read all the samples, and write them to another file. Then I added a function that could do some simple processing before writing the audio. And then I added some more stuff, and some more, and so on.
As much as possible, make early prototypes that actually do some useful subset of the problem you are trying to solve. If your program will have users, give them early prototypes and see what they say. The Software Tools book had an aphorism that 80% of the problem solved now is usually better than 100% solved later.
You may also find that, as the users try out your prototypes, they may discover surprising things. Perhaps what they originally asked for isn't what they really wanted, and you need to drastically redesign. Perhaps once they start using your prototype, they may invent new features that they really want more than some of the other stuff they asked for.
And perhaps you may get a surprise: "Yeah, we told you you could have 8 months to develop it, but now you only get 4." And three of those months are already used up. The left-corner design hopefully means you will deliver something. And it might just be enough.
The opposite school of design would be to think everything through and plan everything. Hold long rounds of meetings, draw diagrams, that sort of thing. That may actually be appropriate in some industries; if that's how they do things where you work, study it and try to figure out if they have a good reason for it. But even if so, you may need to knock out some sort of handy utility for your own convenience, and left-corner design is the way to do that.
steveha
The Elements of Programming Style
on
Clean Code
·
· Score: 5, Informative
I will seek out Clean Code and take a look at it. But I'd like to take this opportunity to plug a classic favorite of mine.
The Elements of Programming Style by Kernighan and Plauger is an old book... so old that all its examples are in either Fortran or PL/I. It doesn't matter. They take examples of code, ruthlessly dissect each one, then rewrite each one; and in every case, their rewritten version is hugely improved. Then they present a rule that encapsulates what they did to improve the example. Their writing is clear, insightful, and entertaining. This is a book that I pull out again and again and re-read.
Right now, eBay is the auction place. Sellers want to sell on eBay because that's where all the buyers are. Buyers want to buy on eBay because that's where all the sellers are.
At some point, both buyers and sellers could get sufficiently upset with eBay as to take their business elsewhere. Are we there yet?
And if we are, what is the most likely competitor?
One more thought: if the current situation is close to the tipping point, then all it would take is one single disturbance to cause a mass exodus from eBay to the new site. For example, if Google were to buy one of eBay's top competitors, and publicly announce Google Auctions, and announce that they would charge less in fees and allow checks and postal money orders... eBay could lose everything in the blink of an eye.
Remember how fast Xfree86 was dumped in favor of X.org? Dissatisfaction with the Xfree86 project was already high, and then they announced one more petty annoying license change, and *boom*, they were done. I wonder if this could happen with eBay.
Slashdot Mirror
When that code needs to be modified to run on a supercomputer instead of a workstation, it is usually converted to pure C or Fortran.
Why is this done? If your Python program is spending almost all of its time inside a C module, why is it worth converting the whole thing to C? Not a flame, I just don't understand. Is it because you run the whole computation over and over in a loop, and the Python overhead starts to be noticeable?
P.S. You are a mathematician... have you seen Sage? It's basically a whole bunch of math libraries, glued together with Python.
http://sagemath.org/
steveha
I wonder if Fortran may eventually be replaced by Python.
A few years ago, when I was first getting into Python, I read an article where a guy from a science research lab talked about his lab's transition from Fortran to Python. Python has some nifty heavy-duty math modules, written in C; and everyone at the lab who tried out the Python stuff strongly preferred it to Fortran.
Since C code is doing all the heavy lifting, it's nice and fast. Since Python is interactive, scientists can use it as a really-powerful desk calculator. And since Python has a clean and friendly syntax, it's easier to write and debug Python programs than Fortran.
I really wish I had saved a copy of that article, or at least its URL. I've tried Google searching for it, and I find many hits on using Python in labs but I haven't found the article.
steveha
Sure, that works. It's still a kludgey workaround that results in you doing part of the compiler's job.
Yet, it's the best I have come up with do to this in Python. I already stipulated that other languages (specifically, Lisp) are easier with respect to this.
Do you have a better way to do it in Python? Or a way to change Python to make this easier... that might actually get accepted into Python? If not, you're just complaining. Which you are of course free to do!
Likewise, if a function is only needed in one place (say, as a callback for a UI event), it's often both more convenient and readable to just stick it inline.
You sound like you think it would be sort of nice to have, which is about where I am on this issue. But some people seem to really care about this.
Python isn't perfect, but in my opinion it's about 95% of the way there and I'm willing to live with the warts (such as the closures hack I documented). I haven't found anything I like better than Python for high-level coding.
steveha
If you need a function that does two things, you can't use lambda anymore. This is not a great hardship as Python allows you to declare inner-scoped functions and you can use that instead,
I'd like to call attention to this point, since everyone but you seems to have missed it.
In Python, lambda expressions are so limited that they are really only useful for trivial purposes. People have asked Guido to improve lambda, but he hasn't. My understanding is that Python's parser somehow makes it hard to parse a multi-line lambda.
The solution is very easy, however.
The purpose of lambda is to create an anonymous function object, often with a closure. Well, in Python, it's trivial to create a function object by declaring a function; you can declare a function anywhere, including in a nested scope, and you can make closures. The only thing you can't do is declare a function with no name. So all you really need is a coding convention, such as: every function I wish was a lambda function will be called "anon_fn".
Here's an example. We want to make a function that returns anonymous function objects that are adders: we will pass in a value x, and we want our adder function objects to remember that x, add it to some number n, and return the result. Once we have make_adder() we can make as many anonymous adder function objects as we wish.
def make_adder(x):
def anon_fn(n):
return x + n
return anon_fn
The above example is trivial enough you could do it with a lambda:
def make_adder(x):
return lambda n: x + n
But I prefer the Python version. In fact, I wouldn't actually use "anon_fn" as the name, I'd probably call the function "adder" and say "return adder".
def make_adder(x):
def adder(n):
return x + n
return adder
make_adder() creates a function object. For a little while, inside make_adder(), this function object is bound to the name "adder", but that binding vanishes when make_adder() returns. The function object is returned, and now it's anonymous.
So, to me the interesting question is not "when will Python get multi-line lambda?" To me, the question is "Why is it so important to be able to make a function that never was bound to a name for even a moment?" This isn't a flame, I'm actually wondering why this seems like such a big deal to so many people.
P.S. As long as I'm beating the lambda horse, I might as well beat the closures horse as well.
People argue that Python doesn't have "true" closures. I think it does, but the way it works may not be obvious.
In Python, variables are really just names that have objects bound to them. If you think of the closure as the name, then closures are broken in Python. If you think of the closure as the object, not so broken.
Here we want to make a function that returns anonymous function objects that are counters. Each counter function object should have an initial value, enclose that value, and when given a number n, add n to the current enclosed value and return the updated value.
# this version seems reasonable but does not work
def make_counter(initial_value):
x = initial_value
def counter(n):
x += n
return x
return counter
What's wrong here? The attempted closure is trying to use a name, "x", where Python needs an object. "x += n" really means "rebind the name 'x' with its previous value plus the value of 'n'". But when you enter counter(), the name x is not bound. Python will enclose the object bound to a name, but doesn't really keep the name binding part.
Over time, all the cool features from proprietary UNIX versions are getting ported to Linux, either directly or by being re-implemented. As Linux becomes more and more acceptable as a replacement, expect to see proprietary UNIX versions start to go away.
If IBM hires a person to work on Linux, that work helps IBM across pretty much their whole product line. If IBM hires a person to work on AIX, that work has much less value now, and will have even less and less value over time as Linux gathers up more of the market. Also, as Linux keeps getting better, it would take more and more work to add similar features to AIX, to try to keep up. Eventually, IBM is going to stop paying for work on AIX at all; they will end-of-life AIX, and just sell Linux.
I don't know for sure about SMIT but Linux does have LVM and various tools to manage it. AIX gurus, how ready is Linux to replace AIX now?
And, are desktop POWER machines going to be available with Linux?
steveha
This web site is perfect for inquisitive teens:
http://scitoys.com/
It's crazy cool. He shows you how to make your own working spectroscope with a box, a CD, two razor blades, and some tape!
The guy who put up that site has written some actual paper books, so you could give one or more of those. Or, just order some magnets and diffraction gratings and such for building the gadgets, from the catalog:
https://www.scitoyscatalog.com/
I really wish I could have had access to that web site when I was 13. Oh, well... at least I have access now!
steveha
Antivirus is one of those things that(at least until actual heuristic scanning that seriously works comes out) leans heavily on having a whole bunch of security guys and worker drones hammering out signature updates all day every day. That isn't something that falls under "The Open Source is strong with this one".
Hmmm, not sure I agree. I have always thought that the open source community could do a great job with antivirus.
The key is to get a large community of people who, when they discover a new virus, contribute their knowledge back to the open source project. And I think this is actually working with ClamAV.
I know that I have submitted my share of viruses... when I get an email offering me a cool new screen saver, and the file is called "screensave.scr.exe", I scan it with ClamAV. If ClamAV doesn't spot anything wrong, I'll submit that file to the ClamAV project.
Usually I submit the file at VirusTotal first, and attach the report to my submission.
ClamAV gets signatures very quickly for new viruses as they appear. The whole signature-based game is a continual game of catchup, though. I agree that heuristic-based scanning would be preferable, but that seems like a hard problem.
steveha
But have you ever considered that the culture might be less violent without a gun culture?
Statistics show that most civilian gun owners are law-abiding and non-violent. I quote Wikipedia:
Most of the violent criminals are already convicted felons, and thus legally barred from owning a gun. I'd say they are part of the criminal culture, not the gun culture.
Have you actually researched the gun culture of the USA? I invite you to read The Gun Culture and Its Enemies. Amazon is offering a used copy for $0.44 plus shipping.
steveha
I've seen you posting on Slashdot, and I usually respect what you write. However, I completely disagree with your conclusions here, and I urge you to think about this some more.
No gun ban has ever succeeded in keeping guns out of the hands of violent criminals. They want guns, so they get them.
Some people will tell you that we could keep guns out of the hands of bad guys if we had a consistent, national, Draconian ban. A War on Guns, if you will. Think for a moment about the War on Drugs. Have we been able to keep crack cocaine out of the hands of users? If not, is it because we don't have a nation-wide Draconian ban on crack cocaine? Now consider that the crack users need to buy more every week. They can get drugs every week, in the face of a nation-wide ban. A violent thug only needs to get a gun once, and then he will carry it for months. (Basically, until he throws it away because cops are after him or something... then he gets another one.)
And, consider how easy it is with a machine shop to make a functioning firearm. Even if all gun makers were driven out of business, and all firearms circulating in society were found and destroyed, and all gun smugglers were caught at the border... all of which I consider impossible, by the way... even then, the bad guys would start making their own guns.
So, it is an axiom with me that no amount of law can keep dangerous weapons out of the hands of the bad guys. What then?
Then, all you can do is disarm the law-abiding. Does that really help?
Now, you mentioned Europe. I urge you to read the book The Samurai, the Mountie, and the Cowboy. I can't summarize a whole book in a few words, but the important takeaway is that violence is much more a cultural thing than a product of laws. England had a low violent crime rate. Then they banned guns. Then they had a low violent crime rate. People will tell you that England banned guns and has a lower crime rate than the USA; that's true, but can you really say it was the ban that caused the lower crime rate? (And in recent years, their violent crime rate has climbed...)
So, I don't believe that you can successfully disarm the bad guys. And I don't believe that you can lower the violent crime rate by passing gun control laws. And history and the law agree that the cops cannot protect everyone; all attempts to sue the police for failing to prevent any crime will fail, even if the police were horribly negligent. Given all this, I strongly oppose any and all efforts to ban firearms in civilian hands.
Also, bad neighborhoods in bad parts of town are where innocent people need firearms the most! I live in a boring, safe suburb, and I don't really need to own a gun. The truly poor people who live in horrible places are in much worse danger than I am. This is one reason I'm opposed to mandatory six-week training classes, expensive licenses, and bans on so-called "Saturday Night Specials". These things do nothing to stop the bad guys, and might keep an innocent citizen from getting a gun in a time of true need. I've read a few stories that curdled my blood.
steveha
There are a whole bunch of "history" tricks, to recall old commands without using the mouse.
When I started college, I studied the shell's man page until I knew them all. Some are so obscure I have forgotten them.
Generally, these involve an '!' character in some way.
Here are a few I use:
!! # run again the last command that was run
!9 # run again the command with history number 9
!v # run again the last command that started with a 'v'
!vi # run the last command that started with "vi"
!?foo? # run the last command that had the string "foo" anywhere in it
diff oldfile newfile
mv !$ !^ # same as "mv newfile oldfile"
# !$ is last arg of previous command, !^ is first arg
ls foo bar baz
rm -f !!* # same as "rm -f foo bar baz"
# !!* repeats all arguments from previous command
There are actually some baroque tricks that recall a previous command and perform a search-and-replace on it, but for anything that complicated I just recall the line and edit it. The baroque tricks would have been pretty darn cool back in the paper teletype days, though.
By the way, the Bash shell can be configured to edit command lines using vi or Emacs commands. I described how to do it in an article I wrote for Linux Journal magazine. It's the last section, "vi or Emacs Mode in the Shell".
http://www.linuxjournal.com/article/8361
Oh, not exactly a history trick, but here's something I use all the time: /tmp/files /tmp/files # edit list to include just the files I want /tmp/files`
ls -1 >
vi
rm `cat
# `cmd` inserts the standard output from cmd into the command line as if you typed it
ls -1 > /tmp/files /tmp/files :%s+.*+mv & /some/directory/path/&+ /tmp/files
vi
# edit list to include just the files I want
# now run this command:
# save file and quit vi
source
This moves the chosen files to "/some/directory/path". The breakdown of the vi command is as so:
: # invoke "ex mode" for search and replace command
% # run the following command on every line of the file
s # do a search and replace
+ # use a '+' for the command delimiter, so I won't have to backslash escape '/' chars in the path
.* # all characters on the line /some/directory/path/& # replace "foo" with "mv foo /some/directory/path/foo"
+ # end the match pattern, begin replace pattern
# & refers to the match pattern, thus all chars on the line
mv &
Takes less time to do it than to explain it!
The above is perhaps overkill if all the files are going to the same place. It's great if you want to send some files one place, some to another, because you can just edit the destinations until it looks right.
steveha
That "nothing has been proven" comment is pretty clearly tongue-in-cheek, just like the "take over the Internet" part.
The video encoding field is crowded with patents, so it's probably impossible to do something like Theora without needing a patent license. But Theora is based on some patented technology (VP3) whose patents have been donated for free use, irrevocably forever. And Theora is free, open source software with a BSD license. If you use the Ogg container format, Theora video, and Vorbis audio, you have a completely free media format.
http://en.wikipedia.org/wiki/Theora
So, you can use Theora for any purpose, without needing to pay royalties, without needing to get permission. That's why it's so funny that Nokia claimed Theora is "proprietary"... I do not think that word means what they think it does.
steveha
Fascinating. Here's a link to the text explaining the car:
http://www.rinspeed.com/pages/cars/squba/pre-squba.htm
This is an actual car, but the Bond version remains sheer fantasy. The Rinspeed's passenger compartment is not pressurized; it's designed to let the water in. According to the above cited text:
The James Bond movie car drove fast on land, and shot wet cement onto the windscreen of a pursuing car, before driving into the sea and then firing a missile to shoot down a helicopter. This is cooler, though, because it actually exists.
steveha
I recommend you just develop a web-based solution using Django. (Or any other web platform of your choice, but Django is the one I have used and I love it. Django makes it easy to do whatever I have needed to do, and I have done some unusual stuff with it.)
Then, you can use any phone with a decent web browser, plus you could use a PDA, a netbook, or anything else. You likely will live in the house for many years, and technology certainly will evolve... a simple web solution is pretty future-proof.
steveha
Google is your friend. There are so many different groups doing Kinetic Sculpture Races... do some searches, find their various web pages. If you live near one of the big races, you might be able to meet people in real life.
But what the heck, I did a few Google searches for you. Here's a few links:
How to Build a Kinetic Racing Sculpture
Links from Kinetic Kingdom
Resources from Kinetic Kensington (in Philadelphia)
Several of those web sites have mailing lists (maybe just for news of the specific race, maybe for general kinetic sculpture discussion, who knows). The first link has a mailto: link from the author ("Elliott Naess") and you could try sending him email to ask his advice.
That http://kinetickingdom.com/ site is a major resource. It's from the oldest kinetic sculpture racing group, in Ferndale, CA.
steveha
Look up Kinetic Sculpture Racing:
http://en.wikipedia.org/wiki/Kinetic_sculpture_race
The guys who do this build wacky and weird human-powered contraptions. These are supposed to go on land, water, sand, and mud, and be only human-powered. (Some vehicles fail in one or more of these categories; but they can avoid being disqualified by providing sufficient bribes to the Kinetic Kops. In plain sight of all onlookers, of course.) Vehicles that can do all of the above, without any "pilots" leaving the vehicle to adjust things, get the "ACE Award" for good engineering.
One of my favorite kinetic sculpture vehicles is a behemoth that carries four people, each of whom provides power to one wheel, and one of whom has the steering wheel and brakes. I have also seen a vehicle that carried eight people, all powering a common drive train.
Anyway, these races have been happening for decades, and you can find the discussion lists where the KSR community discusses where to get parts, how to make things strong and reliable, etc.
For the glory!
steveha
I want it too. When I used to run Windows 98 with Norton Utilities, I had this ("Norton Protected Recycle Bin") and it was slick. See my 6-year-old comment:
http://linux.slashdot.org/comments.pl?sid=39457&cid=4213113
To do this properly, it should be in the file system. (The undelete tool and the expire-old-data tool can and should be in user space, but the data protection features should be in the file system.)
steveha
I think the main thing I see is that the old argument, "Scripting languages are far too slow!" has finally been put to rest.
Well, that was always an oversimplification anyway. Too slow for what, specifically?
Even today, no one would seriously think about writing a video encoder entirely in a dynamic interpreted language. That's a very compute-intensive application and you can't afford the overhead. But how many of us write video encoders? There are many tasks for which the overhead of a dynamic interpreted language is no big deal.
Computers are really fast these days, so you can afford some overhead. If your trivial program runs in 0.6 seconds instead of 0.01 seconds, you may not care about the difference. And if you can write your program in 1/10 the time, you may come out way ahead. (And if the program is a one-off, that only needs to be run a few times, all you really care about is how much of your time it took to write the thing and get it correct.)
steveha
Instead of spending $300 on Bose headphones with active noise cancellation, consider spending $80 to $100 on some headphones that simply block background noise very well.
As I type this, I am listening to music on my Sennheiser HD-280 Pro headphones. According to Sennheiser they provide 30 dB of attenuation for background noise. They simply seal around your ears and passively block noise.
I was talking to a world-class expert on audio stuff one time, and if I recall correctly he was concerned about the noise canceling headphones possibly having a negative effect on your hearing. The active noise cancellation works by measuring the noise, and generating more noise with the phase opposite; if all goes well, the noise and the anti-noise cancel. But the active cancellation is pumping more energy out right next to your ear.
I'm not 100% certain I am remembering correctly what he said. I mean, if you wear the active cancellation headphones and the noise sounds quieter, doesn't that mean that less energy is reaching your ears? I'll have to ask him when I see him again. But that will be too late for this Slashdot discussion.
But you can spend less money to get a simpler device that passively blocks noise, so it seems like the way to go in any event.
By the way, I am quite satisfied with the audio quality of my HD 280 Pro headphones. They are the best headphones I have owned so far.
http://www.headphone.com/products/headphones/sealed-and-noise-canceling/sennheiser-hd-280-pro.php
There are other sealed headphones you can get, including several that are under US$30.
http://www.headphone.com/products/headphones/sealed-and-noise-canceling/
Protect your hearing, folks. You will miss it if you lose it.
steveha
Over time, computers have become more and more powerful. And people always find uses for the additional power. If anyone tells you that we "don't need more power", ignore them.
But the netbooks show that there is a solid niche for people doing basic tasks wherever they want to go. I love going to small local espresso shops, and enjoying some sort of coffee drink while reading Slashdot. Even my OLPC XO is adequate for this, let alone an Atom-powered netbook. If anyone tells you that netbooks "aren't powerful enough", ignore them.
In fact, netbooks are powerful and have huge screens compared to PDAs and phones, and web sites shouldn't ignore PDAs and phones either. All the major web sites need a three screen strategy: their sites should work on TVs for living-room users, the site should work on PCs, and their sites should work on phones.
P.S. I was amused by the example from the summary. Video editing? Oh noes, my netbook is underpowered for video editing. How many people thought otherwise?
steveha
http://www.theinquirer.net/gb/inquirer/news/2008/10/02/norway-standards-members-walk
I was shocked by how excellent the "rough Google translation" was. Unless they had a human clean up the translation a bit, that is amazingly good English prose for a machine translator to emit. (I can't speak for how accurate it is, but it seems plausible enough.)
English is a mess, with lots of irregular usages. How about Norwegian -- is it particularly easy or particularly hard to translate?
steveha
"Do not be afraid to throw it away and start again."
I pretty much agree with most of your points. However, if you have done the left-corner design properly, you should have important building blocks ready to re-use when you do scrap some work and start again.
The audio DSP engine originally had to write the same exact audio format that it originally read; if you fed it 2-channel (stereo) data, you had to get 2-channel data out. When I rewrote it to add the mixer feature, I had to add quite a bit of code. But my code for reading and writing wave audio was perfectly re-usable, as was the actual DSP function code.
steveha
25 years ago I read a book called Software Tools in Pascal. This had a huge, beneficial impact on me.
The most important single thing I learned from that book is something they called "left-corner design". It goes like this:
Find some small part of your project, preferably something at the lowest design layer. Then code it up and implement it. Make it so brain-dead simple that you can spot all the bugs, and fix them. Now, consider some simple way to make it do something more, something else needed for the final product. Then make it do that, and fix it until the new feature is also working perfectly. Iterate.
As a real-life example, I once made an advanced audio DSP (digital signal processing) engine. It started out as a program that could open a wave audio file, read all the samples, and write them to another file. Then I added a function that could do some simple processing before writing the audio. And then I added some more stuff, and some more, and so on.
As much as possible, make early prototypes that actually do some useful subset of the problem you are trying to solve. If your program will have users, give them early prototypes and see what they say. The Software Tools book had an aphorism that 80% of the problem solved now is usually better than 100% solved later.
You may also find that, as the users try out your prototypes, they may discover surprising things. Perhaps what they originally asked for isn't what they really wanted, and you need to drastically redesign. Perhaps once they start using your prototype, they may invent new features that they really want more than some of the other stuff they asked for.
And perhaps you may get a surprise: "Yeah, we told you you could have 8 months to develop it, but now you only get 4." And three of those months are already used up. The left-corner design hopefully means you will deliver something. And it might just be enough.
The opposite school of design would be to think everything through and plan everything. Hold long rounds of meetings, draw diagrams, that sort of thing. That may actually be appropriate in some industries; if that's how they do things where you work, study it and try to figure out if they have a good reason for it. But even if so, you may need to knock out some sort of handy utility for your own convenience, and left-corner design is the way to do that.
steveha
I will seek out Clean Code and take a look at it. But I'd like to take this opportunity to plug a classic favorite of mine.
The Elements of Programming Style by Kernighan and Plauger is an old book... so old that all its examples are in either Fortran or PL/I. It doesn't matter. They take examples of code, ruthlessly dissect each one, then rewrite each one; and in every case, their rewritten version is hugely improved. Then they present a rule that encapsulates what they did to improve the example. Their writing is clear, insightful, and entertaining. This is a book that I pull out again and again and re-read.
http://www.amazon.com/Elements-Programming-Style-Brian-Kernighan/dp/0070342075/ref=sr_1_2?ie=UTF8&s=books&qid=1222277636&sr=1-2
steveha
This reminds me of a gag from Mike Nesmith's Elephant Parts.
"We're not just hoping you're dumb, America... we're banking on it!"
http://www.youtube.com/watch?v=pEPxc3RW4js
steveha
Right now, eBay is the auction place. Sellers want to sell on eBay because that's where all the buyers are. Buyers want to buy on eBay because that's where all the sellers are.
At some point, both buyers and sellers could get sufficiently upset with eBay as to take their business elsewhere. Are we there yet?
And if we are, what is the most likely competitor?
http://online-auction-sites.toptenreviews.com/
One more thought: if the current situation is close to the tipping point, then all it would take is one single disturbance to cause a mass exodus from eBay to the new site. For example, if Google were to buy one of eBay's top competitors, and publicly announce Google Auctions, and announce that they would charge less in fees and allow checks and postal money orders... eBay could lose everything in the blink of an eye.
Remember how fast Xfree86 was dumped in favor of X.org? Dissatisfaction with the Xfree86 project was already high, and then they announced one more petty annoying license change, and *boom*, they were done. I wonder if this could happen with eBay.
http://www.internetnews.com/dev-news/article.php/3338031
steveha