Which really, really irritates me. People talk about red light cameras and speed traps as if they were some evil violation of the constitution. When you point out that speeders and red light racers kill people, they spout conspiracy theories about doctored cameras and shortened yellow lights.
Lots of things kill people. Indeed lots of bad driving habits kill people, like driving too close to the car in front... but the police almost never look at those. But yet they spend $large_amount hiding at the bottom of hills to catch you doing 20% over the limit on an empty interstate. But, sure, you tell yourself that it's all "for the children" if that makes you happier.
Re:I am afraid, there is lack of direction for Rub
on
Ruby 1.9.1 Released
·
· Score: 1
Blocks are also more concise.
C/Java/Whatever doesn't require the newlines... if you prefer non-newlines then:
List<Person> children = new List<Person>(); for (Person p : people) if (p.age < 18) children.add(p);
...now the only real difference is the fact you are declaring what types are in the list, something ruby/python/perl can't do and is one of my biggest annoyances with python (don't do ruby).
Because it's been a long time since my view of what a bug is would meet that definition. Almost all bug trackers have priority/severity/whatever markings... so that the very small number of bugs that fit into your category can stand out. On all but the very worst software those are "always empty" as you describe, but they account for probably somewhere between 0.01% and 1% of the actual "bugs".
To a user everything that doesn't work as they want it to is a bug, some of the things you can easily argue are technically features so you can easily argue are "severity high bugs" or whatever. But what about "Blah uses too much memory/CPU, when I feed it 10,000x the usual input", this could very well be argued to be either a bug (if scaling was expected) or a feature (if it was known it wouldn't scale past 10x usual input). And as I said, IME users just don't care if it doesn't solve all their problems it's buggy (and they wouldn't be using it if they didn't expect it to solve their problems). What about "software uses sha1 and we now need it to use sha256", that is obviously a new feature but in many situations is considered a medium to high security bug.
Sorry, but "fix all submitted bugs" is just not possible, unless you are using a different definition of some of those words than everyone else. Or more likely you get a very limited set of bug reports, for whatever reason.
For a start the difference between a "bug" and a "feature" is non-existant to any normal user, so your code would have to be Skynet to not have any bugs. Also almost all users will have no idea how much work a bug entails, so you tend to have reports of "make application play alert sound when Foo happens, by default" (and the opposite default) along with "make X work (with the same features) on Windows, X, text and over serial console".
What do you mean by "adversely"? If Coke make me think one of their drinks is better than one of Pepsi's, does it matter if I change my opinion or not? Where is the adversity?
People want to say video games alter people to the point where they are normal one day and killers the next
Some do, yeh, but not many people here... and I doubt that is what the GP post was arguing against. Mostly intelligent people often say things like "advertising doesn't affect me", which I don't take to mean "Yeh, they can get me to buy their product but I don't instantly start worshipping the company and give them all my money" I take it to mean "in a double blind test nothing different would happen"... and in the general case I find it really hard to take that on faith.
Humanity produces but ~2% of it, the rest is produced naturally by volcanoes, forest fires and such.
You probably want to stop reading Fox and Rush, esp. 14 years after their lies have been debunked:
July/August 1994
The Way Things Aren't
Rush Limbaugh Debates Reality
It's been done; Klik is your answer. From the Wikipedia article:
Why stop at one non-std. package manager. There is also autopackage, zero-install and nixOS... what is the common theme of all of these. They are all complete failures, all the main distros. still use either dpkg+apt or rpm+yum (or zypper) and recommend you don't use anything else. Instally random crap from random places on the network has always been a bad idea, and not a feature you want to port from Windows/MacOSX, trying to "work around" that by having 1,000 repos. or whatever is still the same bad idea dressed differently.
Without creating a third mainstream package manger (and based on a bad idea, at that), the only viable "cross platform" option is PackageKit... except that only solves the trivial problem badly, what the actual package is called that provides what you want is still specific to each distro.
So, yeh, you could have something in firefox in Ubuntu that converts blah:firefox into "apt-get install firefox" (or s/firefox/whatever/)... but it would then only work in Ubuntu, and even then maybe only in specific versions of same. Or you could just admit that having a native application is probably a better idea than random web page links.
It never was clear to me that, at any time, Linux was actually telling the drives to copy data directly from one disk to any other without the kernel in the middle.
IIRC it was proposed on lkml, however it would still need to use the SCSI bus which is where the majority of the time is spent anyway. Also nothing else had tried to do that, so everyone was worried that it'd be turned on and would have weird failure cases (which would be very bad.
Saying that a condo owner can't place certain things ON THE PROPERTY THAT THEY DON'T own is fine by me too
But you buy a condo, and they do place restirctions on what you can/can't do with that thing you bought. As does the local and state govt. when you buy a house.
I consider many/most Zoning regulations just as unjust.
Sure, and I know of people who consider restrictions placed on them by building codes "unjust" too (including wiring their heat into the box your rentees pay for -- hey, they "own" the building). People consider lots of things "unjust" when it stops them screwing other people over, not so much when it stops other people screwing them over. HOA are the same thing, but much better as you can easily just not buy a house under one.
My stance is that people should be able to do anything they like on their own property.
Except Zoning and Building codes already restrict what you can do, and in most cases for good reason (although some again, are "beautification" rules like "no signs on a business"). HOA are just a more localized version of that, and HOAs are much less problematic because it's much easier to just find a different house that isn't under a HOA you don't like.
The other way to look at it is: HOA => Condos => Rental. You'd probably agree that if you are renting the owner should have the final say over what you can do, Condos use the same argument that you are buying and renting and HOA as the final part where you buy but still have rent like restrictions.
To just say it bluntly: almost all personality tests are completely unreliable, the best ones are at most somewhat reliable. This means that most people, when they do the same personality test some time later, score differently. This is a terrible problem, as the point of these tests often is to measure a personality, something presumably stable over time.
Sure, and that's almost certainly true of any "real" exam too. That doesn't mean employers are going to stop asking for a HS diploma or Degree. I agree some of it is "we need to do something, this is something, we need this", but some of it is also that anything which helps reduce the number of possible candidates from 1,000 to 20 solves the time problem... and you don't care if there are false positives.
The whitespace issue is a red-herring: most people get used to it quickly
You have data to back this up, I assume? I'm not certain that the majority of people I know using python hate the whitespace retardedness, but I know it's not a "red-herring" and it's far from 90%+ of people who don't hate it (people either don't care or hate it, with a very few who thinks it's a great idea... IME).
There's two real-world problems with it: copy-and-paste and generating Python code.
Err... and let's not forget that if you write a goddamn doc comment at the wrong level your program crashes. Or that python code tends to be much more crammed together because everyone is afraid of "lining up" blocks. Or that minimal patching is often impossible. Or, hell, that it's impossible to just position the cursor and start typing (nevermind, as you say, the fact moving code with cut/paste doesn't work). Or that half the people trying to use the editors they've been working with for years can't work the same way with python (for instance using TAB in xemacs to "reset indentation to correct level" just can't work in python)... or that the other half tend to use a special python editor, just so they can work around the invisible syntax.
Then there's the bits that seem to be there just to annoy developers, like that fact that it's almost impossible to write functions in ipython to test something out because the whitespace will almost certain screw you over... so instead you have to spend 2-10x as long writing the entire thing out in a script (along with all the state of where you are).
Then as you pointed out there's the awesome insanity of TABs vs. SPACEs, so things that look different aren't! Luckily you can pass -tt, unluckily that affects all the modules you import (and what they import)... hahahaha!
And then, just when you are deciding between stabbing your eyes out, breaking the spacebar on GvR's keyboard or patching "from __future__ import braces" to actually work... someone posts saying that "it's not a problem really".
Meanwhile, in academia, they would be using "research toys" like trucks, trains, airliners and huge ships to transport 100-ton objects (or packets of smaller objects) between campuses and research stations.
Terrible analogy, those trucks are interoperable if companyA moves to using trucks/trains/etc. then they can use those to ship to companyB. If any company moved to be ipv6 only they'd have the same effect as if they powered down their data center.
I currently pay ~$100 for a/29, and given I'm not a business I'd at least consider moving to ipv6 for economical reasons... except the last time I asked my ISP they said they don't offer it as it'd be more expensive. And I'm also pretty sure it wouldn't "just work" even on outgoing connections, like playing on my PS3.
IPv6 will happen when they make the pain of moving less than the pain of not moving. One obvious way is to make it fully backwards compatible for 5 years, or so, and have it be at least a little cheaper/better than using IPv4 only. The next most obvious way (and my guess for what will happen) is that IPv4 will hit a wall that will be massively painful, at which point the POS that is current IPv6 will be the lesser of two evils.
The letter was informing me that they had decided to change the license terms on their hardware - after my purchase, that signing up for their service was "mandatory", and that if I did not do so within X number of days or receiving my device they would CHARGE MY CREDIT CARD.
Personally I would have just ignored this, and if they charged me I'd have called my CC company and said it was an unauthorized charge. I also am not optimistic enough to assume that whatever happened wouldn't have happened at Best Buy or any other large chain.
Don't do it. Really -- don't do write a book. [...] wrote a book (http://valdyas.org/python/index.html). Total income for about a year of working every evening and all weekends, ~ $400.
Now to be fair to you, I've seen this advice from a significant number of people who have written technical books (it does not pay minimum wage, do not do it for the money). But I will note that your book is "GUI programming with Python and Qt" so the obvious retort is "$400... wow all the people who wanted it, bought it!"
I don't believe this is correct. If you explicitly and validly release a copyrighted work into public domain, then it leaves your control permanently. Everyone has the exact same rights to the work as yourself.
Depends on which laws (countries) you are talking about, SO has a pretty good question/answer about it:
Not quite sure what you mean by this. GC or the fact that C++ uses less memory than Java.
In the link you gave to the new shootout, it has three "comparison type" boxes: 1) CPU. 2) MEM. 3) LOC. Java does well if you just turn on the first box, I explained what happened as you turned the other two boxes on.
And, unlike the other poster, I'm not implying that Java isn't "fast enough" just that "faster than C or C++" is misleading at best.
This is called optimization and we do it using various types of tools.
This isn't quite the same thing, for instance I've seen code that refused a real string type in C (for "performance"/"memory") and thus. strcmp() was at the top of it's "cumlative time in functions" list in certain paths. I've also seen code which had all it's used memory as a mmap() of a single file (basically a custom DB) with all it's data being pointers into that memory, which made it pretty unmaintainable -- but argubly the speed/memory saving was worth it.
One of the joys of C is that programmers using it tend to make those kinds of choices more often than their counterparts using Java or Python. But it does cast somewhat of a shadow on the shootout as a reference point between the performance of Java and C (as they don't accept solutions in that style).
As the benchmark here shows, java is very competitive with C++.
Sure, until you add the memory used back in... and then it falls to mid way between C and Python. Of course you then add in "size of code" (read: programmer time) and it's about par with Python... and it's much easier to hook C code into python for the bits that need to be fast.
And the new "shootout" specifically asks for the "simplest" versions in each language, so I'd say it significantly underreports C/C++ code's speed compared to other things (I see a lot of "in the wild" C code that does really weird stuff so it can, supposedly, get a tiny speed/mem imporovement).
But, sure, feel free to take note of only the bits of info. that makes your point look credible and be shocked that everyone doesn't automatically agree.
Child Porn task forces are highly focused units with very specific and narrowly defined missions. They are NOT out there after Hentai viewers, for instance. They are only after REAL Child porn, both it's purveyors and consumers.
Geez...rather that all this unionization, and all...why don't we go back more to where everyone IS more of an indy contractor, and let them negotiate for pay and benefits. That way..the cream will rise to the top, and get paid top rate....the lesser ones...well, may have to pick another field to work in?
The last time I looked the skill "negotiating for wages/benifits" did not corrolate to the skill "software developer"/etc.... so, yes, the "cream" will rise but it's the people with the first skill and not the people with the second who will be considered the cream. And while I've seen some good "software contractors", the majority fit this rule.
Personally I think most software developers vote against unions for the same reason they vote republican, they are idiots who are under the delusion they are doing much better than they are.
Everyone knows the market is going to be way up in a few years because it is currently highly undervalued...
You sound like the prognosticators in 1929. But it took 22 long years for the Dow to surpass its pre-depression highs. Don't commit the same sin of hubris that got us here in the first place.
The DOW as it exists now didn't exist in 1922, but anyway what I assume the GP was trying to say is that in "a few years" the DOW will be significantly higher than 8,000. You are arguing that it might well take "a long time" for it to surpass 14,000. You might both be correct.
All recursion can be done by hand, using loops etc. see
this google result for "Ackerman function non-recursive". This is approach can be much faster (and is never slower), but often much uglier than the "simple" recursive version. It also doesn't have the weird edge case failures though (for instance blowing the stack in C/python/etc. is basically unrecoverable).
Re:But does it run on .... shit that does not work
on
Fedora 10 Released
·
· Score: 1
Anything specific you don't like? Not that/. is a bugtracking system, but...
Slashdot Mirror
That's funny, because os.fork() etc. work fine on my version of python.
Lots of things kill people. Indeed lots of bad driving habits kill people, like driving too close to the car in front ... but the police almost never look at those. But yet they spend $large_amount hiding at the bottom of hills to catch you doing 20% over the limit on an empty interstate. But, sure, you tell yourself that it's all "for the children" if that makes you happier.
C/Java/Whatever doesn't require the newlines ... if you prefer non-newlines then:
...now the only real difference is the fact you are declaring what types are in the list, something ruby/python/perl can't do and is one of my biggest annoyances with python (don't do ruby).
Because it's been a long time since my view of what a bug is would meet that definition. Almost all bug trackers have priority/severity/whatever markings ... so that the very small number of bugs that fit into your category can stand out. On all but the very worst software those are "always empty" as you describe, but they account for probably somewhere between 0.01% and 1% of the actual "bugs".
To a user everything that doesn't work as they want it to is a bug, some of the things you can easily argue are technically features so you can easily argue are "severity high bugs" or whatever. But what about "Blah uses too much memory/CPU, when I feed it 10,000x the usual input", this could very well be argued to be either a bug (if scaling was expected) or a feature (if it was known it wouldn't scale past 10x usual input). And as I said, IME users just don't care if it doesn't solve all their problems it's buggy (and they wouldn't be using it if they didn't expect it to solve their problems). What about "software uses sha1 and we now need it to use sha256", that is obviously a new feature but in many situations is considered a medium to high security bug.
Sorry, but "fix all submitted bugs" is just not possible, unless you are using a different definition of some of those words than everyone else. Or more likely you get a very limited set of bug reports, for whatever reason.
For a start the difference between a "bug" and a "feature" is non-existant to any normal user, so your code would have to be Skynet to not have any bugs. Also almost all users will have no idea how much work a bug entails, so you tend to have reports of "make application play alert sound when Foo happens, by default" (and the opposite default) along with "make X work (with the same features) on Windows, X, text and over serial console".
What do you mean by "adversely"? If Coke make me think one of their drinks is better than one of Pepsi's, does it matter if I change my opinion or not? Where is the adversity?
Some do, yeh, but not many people here ... and I doubt that is what the GP post was arguing against. Mostly intelligent people often say things like "advertising doesn't affect me", which I don't take to mean "Yeh, they can get me to buy their product but I don't instantly start worshipping the company and give them all my money" I take it to mean "in a double blind test nothing different would happen" ... and in the general case I find it really hard to take that on faith.
You probably want to stop reading Fox and Rush, esp. 14 years after their lies have been debunked: July/August 1994
The Way Things Aren't
Rush Limbaugh Debates Reality
Why stop at one non-std. package manager. There is also autopackage, zero-install and nixOS ... what is the common theme of all of these. They are all complete failures, all the main distros. still use either dpkg+apt or rpm+yum (or zypper) and recommend you don't use anything else. Instally random crap from random places on the network has always been a bad idea, and not a feature you want to port from Windows/MacOSX, trying to "work around" that by having 1,000 repos. or whatever is still the same bad idea dressed differently.
Without creating a third mainstream package manger (and based on a bad idea, at that), the only viable "cross platform" option is PackageKit ... except that only solves the trivial problem badly, what the actual package is called that provides what you want is still specific to each distro.
So, yeh, you could have something in firefox in Ubuntu that converts blah:firefox into "apt-get install firefox" (or s/firefox/whatever/) ... but it would then only work in Ubuntu, and even then maybe only in specific versions of same. Or you could just admit that having a native application is probably a better idea than random web page links.
IIRC it was proposed on lkml, however it would still need to use the SCSI bus which is where the majority of the time is spent anyway. Also nothing else had tried to do that, so everyone was worried that it'd be turned on and would have weird failure cases (which would be very bad.
But you buy a condo, and they do place restirctions on what you can/can't do with that thing you bought. As does the local and state govt. when you buy a house.
Sure, and I know of people who consider restrictions placed on them by building codes "unjust" too (including wiring their heat into the box your rentees pay for -- hey, they "own" the building). People consider lots of things "unjust" when it stops them screwing other people over, not so much when it stops other people screwing them over. HOA are the same thing, but much better as you can easily just not buy a house under one.
Except Zoning and Building codes already restrict what you can do, and in most cases for good reason (although some again, are "beautification" rules like "no signs on a business"). HOA are just a more localized version of that, and HOAs are much less problematic because it's much easier to just find a different house that isn't under a HOA you don't like.
The other way to look at it is: HOA => Condos => Rental. You'd probably agree that if you are renting the owner should have the final say over what you can do, Condos use the same argument that you are buying and renting and HOA as the final part where you buy but still have rent like restrictions.
Sure, and that's almost certainly true of any "real" exam too. That doesn't mean employers are going to stop asking for a HS diploma or Degree. I agree some of it is "we need to do something, this is something, we need this", but some of it is also that anything which helps reduce the number of possible candidates from 1,000 to 20 solves the time problem ... and you don't care if there are false positives.
You have data to back this up, I assume? I'm not certain that the majority of people I know using python hate the whitespace retardedness, but I know it's not a "red-herring" and it's far from 90%+ of people who don't hate it (people either don't care or hate it, with a very few who thinks it's a great idea ... IME).
Err ... and let's not forget that if you write a goddamn doc comment at the wrong level your program crashes. Or that python code tends to be much more crammed together because everyone is afraid of "lining up" blocks. Or that minimal patching is often impossible. Or, hell, that it's impossible to just position the cursor and start typing (nevermind, as you say, the fact moving code with cut/paste doesn't work). Or that half the people trying to use the editors they've been working with for years can't work the same way with python (for instance using TAB in xemacs to "reset indentation to correct level" just can't work in python) ... or that the other half tend to use a special python editor, just so they can work around the invisible syntax.
Then there's the bits that seem to be there just to annoy developers, like that fact that it's almost impossible to write functions in ipython to test something out because the whitespace will almost certain screw you over ... so instead you have to spend 2-10x as long writing the entire thing out in a script (along with all the state of where you are).
Then as you pointed out there's the awesome insanity of TABs vs. SPACEs, so things that look different aren't! Luckily you can pass -tt, unluckily that affects all the modules you import (and what they import) ... hahahaha!
And then, just when you are deciding between stabbing your eyes out, breaking the spacebar on GvR's keyboard or patching "from __future__ import braces" to actually work ... someone posts saying that "it's not a problem really".
Terrible analogy, those trucks are interoperable if companyA moves to using trucks/trains/etc. then they can use those to ship to companyB. If any company moved to be ipv6 only they'd have the same effect as if they powered down their data center.
I currently pay ~$100 for a /29, and given I'm not a business I'd at least consider moving to ipv6 for economical reasons ... except the last time I asked my ISP they said they don't offer it as it'd be more expensive. And I'm also pretty sure it wouldn't "just work" even on outgoing connections, like playing on my PS3.
IPv6 will happen when they make the pain of moving less than the pain of not moving. One obvious way is to make it fully backwards compatible for 5 years, or so, and have it be at least a little cheaper/better than using IPv4 only. The next most obvious way (and my guess for what will happen) is that IPv4 will hit a wall that will be massively painful, at which point the POS that is current IPv6 will be the lesser of two evils.
Personally I would have just ignored this, and if they charged me I'd have called my CC company and said it was an unauthorized charge. I also am not optimistic enough to assume that whatever happened wouldn't have happened at Best Buy or any other large chain.
Now to be fair to you, I've seen this advice from a significant number of people who have written technical books (it does not pay minimum wage, do not do it for the money). But I will note that your book is "GUI programming with Python and Qt" so the obvious retort is "$400 ... wow all the people who wanted it, bought it!"
Depends on which laws (countries) you are talking about, SO has a pretty good question/answer about it:
http://stackoverflow.com/questions/219742/open-source-why-not-release-into-public-domain
In the link you gave to the new shootout, it has three "comparison type" boxes: 1) CPU. 2) MEM. 3) LOC. Java does well if you just turn on the first box, I explained what happened as you turned the other two boxes on.
And, unlike the other poster, I'm not implying that Java isn't "fast enough" just that "faster than C or C++" is misleading at best.
This isn't quite the same thing, for instance I've seen code that refused a real string type in C (for "performance"/"memory") and thus. strcmp() was at the top of it's "cumlative time in functions" list in certain paths. I've also seen code which had all it's used memory as a mmap() of a single file (basically a custom DB) with all it's data being pointers into that memory, which made it pretty unmaintainable -- but argubly the speed/memory saving was worth it.
One of the joys of C is that programmers using it tend to make those kinds of choices more often than their counterparts using Java or Python. But it does cast somewhat of a shadow on the shootout as a reference point between the performance of Java and C (as they don't accept solutions in that style).
Sure, until you add the memory used back in ... and then it falls to mid way between C and Python. Of course you then add in "size of code" (read: programmer time) and it's about par with Python ... and it's much easier to hook C code into python for the bits that need to be fast.
And the new "shootout" specifically asks for the "simplest" versions in each language, so I'd say it significantly underreports C/C++ code's speed compared to other things (I see a lot of "in the wild" C code that does really weird stuff so it can, supposedly, get a tiny speed/mem imporovement).
But, sure, feel free to take note of only the bits of info. that makes your point look credible and be shocked that everyone doesn't automatically agree.
Sure, the most recent case of narrowly defined missions being: http://news.bbc.co.uk/1/hi/scotland/7787845.stm (although it's in the UK it's the same idea) the previous big wtf I saw from the us being http://www.newarkadvocate.com/article/20081008/NEWS01/810080302 (actually there have been a bunch of cases like that -- no suprise taking common sense out of the equation to "save the children" screws everything up).
The last time I looked the skill "negotiating for wages/benifits" did not corrolate to the skill "software developer"/etc. ... so, yes, the "cream" will rise but it's the people with the first skill and not the people with the second who will be considered the cream. And while I've seen some good "software contractors", the majority fit this rule.
Personally I think most software developers vote against unions for the same reason they vote republican, they are idiots who are under the delusion they are doing much better than they are.
The DOW as it exists now didn't exist in 1922, but anyway what I assume the GP was trying to say is that in "a few years" the DOW will be significantly higher than 8,000. You are arguing that it might well take "a long time" for it to surpass 14,000. You might both be correct.
You're right, all code should look equal good ... then it's much harder to tell if a moron wrote it, or someone with some taste.
Besides it's much more fun to have your program crash because someone accidentally got the indentation for a doc string wrong.
All recursion can be done by hand, using loops etc. see this google result for "Ackerman function non-recursive". This is approach can be much faster (and is never slower), but often much uglier than the "simple" recursive version. It also doesn't have the weird edge case failures though (for instance blowing the stack in C/python/etc. is basically unrecoverable).
Anything specific you don't like? Not that /. is a bugtracking system, but...