I think performance is an overrated concern for most apps.
I disagree with that. If you're a professional, time is money, and performance always matters.
What doesn't always matter is run-time performance. Programmer performance is often just as, or more, important. A program that takes three hours to write and ten minutes to run is inferior to one that takes an hour to write and half an hour to run, if it's only going to be run once.
Even if it's not going to be run only once, you have to take into account the cost of deployment and maintenance (bug fixing and feature enhancement). If it takes 10 minutes to fix a bug as opposed to an hour, that's a real benefit which you can measure in dollars.
This is one link that I could find. Granted, it's from a popular science TV programme in Australia, and it's mostly about lane changing, but it has references at the bottom.
The real problem is that zoning laws force us to live so far from work, and that those same laws allow/force us into high density areas where traffic will increase.
I don't know where "us" is, but from the sound of it, the real problem is that wherever you live doesn't have decent mass transit. High density is no problem if one vehicle can carry a few hundred people.
You believe that the same amount of cars behind you will make it through the intersection whether you travel quickly or slowly to the next intersection? Really?
On average, the same number of cars will make it through the following intersection.
There have been many tests of this. In heavy traffic, cars which accellerate/decellerate/weave average a roughly 1% speed increase (e.g. 31 km/h vs 30 km/h) at a cost of roughly 50% more fuel used. So you're right, the "it doesn't cost them any time at all" claim is only 99% correct.
Anyone who tells you that you can't go 60MPH on the interstate when there's an inch of snow on the ground are either overly cautious or stupid.
There are sometimes good excuses for being cautious. The driver in question could be relatively new at this driving gig, or from somewhere that doesn't have snow, or from somewhere that has black ice under snow.
As much as I hate software patents, I concede that there's an argument for making truly innovative algorithms (e.g. RSA) patentable. The problem is that over 90% of software patents are of the form of "we do X-prior-art and Y-prior-art... only together!" or "we do X-prior-art... only with the Internet!"
Clearly people like the sound of distorted, noisy playback. I just for the life of me can't understand why.
Because it sounds better than distorted, clean playback.
Hell, in digital mobile phones, they add noise to what you hear to cover up aliasing caused by the sampling/compression algorithm. The noise sounds better than the aliasing.
With "real" generics the system has two choices: either generate lots of bloated specific instances of the code, or add type-checking at runtime.
There's a third choice: Design the language from the ground-up with generics in mind, and don't try hacking it on later to a model where it doesn't work properly.
Excuse me.. but you are trying to justify to me that it is better to write certain forms of software in functional languages, but you're not willing to show me the code that is written in that way so that I can personally evaluate whether or not the code is "better".
First off, I didn't make the original claim.
Secondly, I agree with you that if you can't see the source code, it's not science.
But thirdly, I still think you're being unfair. You don't need to see the plans for a Boeing aircraft and a Tupolev aircraft to know that you're better off riding on Boeing. All you need to do is look at the crash record. Similarly, there are plenty of observables: Robustness, speed of fixing bugs, time-to-market etc, which, taken together, form a pretty good picture of how good code is.
[...] or may be so convulted and unmaintainable that it doesn't matter how well it stacks up in benchmarks.
If you want to know how maintainable code is, submit a problem report. If it's not fixed in a timely manner, that doesn't prove anything (it could be the code, or could be the team). If it is, then the code has to be at least pretty maintainable.
Look, I don't think I'm asking something too unreasonable here.
By limiting to open source only, I think you are.
Let's take the classic example of a difficult multi-threading problem: a DBMS. It's not completely separable, like an FTP server, and it's not effectively serialised, like an X11 server.
Have people written highly-threaded DBMSes in functional languages? Sure. Are they open source? Probably not. There are a few decent open source DBMSes, so there's really no itch to scratch. (So why was Mnesia written? For telecommunications applications, like telephone exchanges; not exactly a haven for open source.)
Proprietary software doesn't count.. why? Cause no-one can see the benefits of using a functional language except the priveleged few who have access to the source code.
And just how many open source applications that rely on parallelisation (and we won't count any kind of parallelisation where there's mostly no coordination required between threads/processes, such as Apache) have been written recently?
The problem as I see it is that I have not seen any arguments for human rights that (a) logically includes all humans (even if we are talking just able bodied and minded ones) and (b) clearly exclude the great apes.
My problem with it is that even able-bodied and able-minded great apes cannot (as far as I know) respect the "rights" of other animals.
You're not allowed to lock me up, torture me and kill me for my meat & skin just because I don't pay taxes or send my offspring (heh) to school.
I'm not, but there are some animals that would have no problem with detaining you, torturing you (well, you'd consider it torture) and killing you for your meat. And they'd have less problem with doing that for another animal. Have you ever seen a cat with a bird?
If animals have rights, then humans are not the only animal in the world that should respect the rights of other animals.
Not in our society. Certain rights are only granted if certain responsibilities are upheld, but even our most despised criminals are granted the right to food, shelter, freedom from torture and so on.
Apart from freedom from torture, most citizens in the US don't have those rights. Most don't even have the right to basic health care, which is something that the rest of the developed world consistently finds astonishing.
The right to food and shelter is not a "right" afforded to criminals, but a duty of the state. If you want to incarcerate them, you have to look after them.
And perhaps most disturbingly, in many places in the US, criminals do not have the right to vote, even after they are released. So much for "no taxation without representation".
However, most rights still have responsibilities. Freedom of speech, surely one of the most basic of basic rights in a democracy, does not include the freedom to engage in slander or libel. That's an associated responsibility.
For the record, I'm specifically talking about a general computer science education, with maths. Now read on...
Intelligence of game bots needs even less math.
No, it doesn't, but how can you make an informed choice between two path-finding algorithms? No, you don't have to have seen and analysed them both before, but you at least need to be able to do the analysis.
As for parallel programming (BTW, one of my favorite languages is Erlang) - you don't need math at all for it. You do need a good knowledge of CS, but it's not university math.
Again, sure. But consider the next generation of lock-free programming. To understand that, you need to be able to reason through cases, understand transactions, and have a pretty thorough knowledge of how the different levels computer cache (including store buffers) work.
What I'm saying is that a specific "games-only" web qualification is no replacement for a decent general computer science education. To be a modern software developer, be it in games or whatever, you need to know how to learn and adapt, far more than you need to know specific industry-specific factoids. And any company that values the latter over the former is not a company that you want to work for.
And if you want to implement AI? What if you're working for a company that isn't interested in last years' games, but next years', which might feature fluid simulation or something? And what about the next generation of multi-core CPUs where locking will be so expensive that you'll need to start using lock-free programming?
It might just be me, but I would not hire anyone who deliberately learned as little as possible, to do only just what was required to do last years' job.
The best favour you can do yourself is get a general education in computer science, with a thorough grounding in software engineering and pure/applied maths.
You're thinking like a slashdotter. Think like an open source businessperson for a moment, instead, and you'll see what's probably the most compelling reason, which is Checkbox completeness.
I can sell you a solution with the "yes, your Windows admins don't need to be retrained" checkbox checked. I can do that without paying the Redmond tax, and like all other open source, I can diagnose and fix problems by myself without being at the mercy of someone else's support cycle, assuming they want to fix your problem at all.
One example is skip lists, as already mentioned. Another common use (yes, I said common use; these things are all over software) is where you have data structures which logically need to be members of more than one set.
Looking at sched.h in my Linux 2.6 kernel source, for example, I see that a task_struct is threaded with no less than six doubly-linked and three singly-linked lists. That's an eleventy-linked data structure, for those keeping count, which makes your measly patented triply-linked list look positively sparse.
Slashdot Mirror
I disagree with that. If you're a professional, time is money, and performance always matters.
What doesn't always matter is run-time performance. Programmer performance is often just as, or more, important. A program that takes three hours to write and ten minutes to run is inferior to one that takes an hour to write and half an hour to run, if it's only going to be run once.
Even if it's not going to be run only once, you have to take into account the cost of deployment and maintenance (bug fixing and feature enhancement). If it takes 10 minutes to fix a bug as opposed to an hour, that's a real benefit which you can measure in dollars.
The same way SCO did, of course.
You did pay your SCO Danegeld, right?
This is one link that I could find. Granted, it's from a popular science TV programme in Australia, and it's mostly about lane changing, but it has references at the bottom.
I don't know where "us" is, but from the sound of it, the real problem is that wherever you live doesn't have decent mass transit. High density is no problem if one vehicle can carry a few hundred people.
On average, the same number of cars will make it through the following intersection.
There have been many tests of this. In heavy traffic, cars which accellerate/decellerate/weave average a roughly 1% speed increase (e.g. 31 km/h vs 30 km/h) at a cost of roughly 50% more fuel used. So you're right, the "it doesn't cost them any time at all" claim is only 99% correct.
I can see it now: "Segway Parking Only"
There are sometimes good excuses for being cautious. The driver in question could be relatively new at this driving gig, or from somewhere that doesn't have snow, or from somewhere that has black ice under snow.
Exactly.
As much as I hate software patents, I concede that there's an argument for making truly innovative algorithms (e.g. RSA) patentable. The problem is that over 90% of software patents are of the form of "we do X-prior-art and Y-prior-art... only together!" or "we do X-prior-art... only with the Internet!"
On Unix, fair scheduling means that all tasks get equally shitty service.
Because it sounds better than distorted, clean playback.
Hell, in digital mobile phones, they add noise to what you hear to cover up aliasing caused by the sampling/compression algorithm. The noise sounds better than the aliasing.
There's a third choice: Design the language from the ground-up with generics in mind, and don't try hacking it on later to a model where it doesn't work properly.
First off, I didn't make the original claim.
Secondly, I agree with you that if you can't see the source code, it's not science.
But thirdly, I still think you're being unfair. You don't need to see the plans for a Boeing aircraft and a Tupolev aircraft to know that you're better off riding on Boeing. All you need to do is look at the crash record. Similarly, there are plenty of observables: Robustness, speed of fixing bugs, time-to-market etc, which, taken together, form a pretty good picture of how good code is.
If you want to know how maintainable code is, submit a problem report. If it's not fixed in a timely manner, that doesn't prove anything (it could be the code, or could be the team). If it is, then the code has to be at least pretty maintainable.
By limiting to open source only, I think you are.
Let's take the classic example of a difficult multi-threading problem: a DBMS. It's not completely separable, like an FTP server, and it's not effectively serialised, like an X11 server.
Have people written highly-threaded DBMSes in functional languages? Sure. Are they open source? Probably not. There are a few decent open source DBMSes, so there's really no itch to scratch. (So why was Mnesia written? For telecommunications applications, like telephone exchanges; not exactly a haven for open source.)
And just how many open source applications that rely on parallelisation (and we won't count any kind of parallelisation where there's mostly no coordination required between threads/processes, such as Apache) have been written recently?
My problem with it is that even able-bodied and able-minded great apes cannot (as far as I know) respect the "rights" of other animals.
An excellent question. How do you get a cat to respect the rights of a mouse?
I'm not, but there are some animals that would have no problem with detaining you, torturing you (well, you'd consider it torture) and killing you for your meat. And they'd have less problem with doing that for another animal. Have you ever seen a cat with a bird?
If animals have rights, then humans are not the only animal in the world that should respect the rights of other animals.
Apart from freedom from torture, most citizens in the US don't have those rights. Most don't even have the right to basic health care, which is something that the rest of the developed world consistently finds astonishing.
The right to food and shelter is not a "right" afforded to criminals, but a duty of the state. If you want to incarcerate them, you have to look after them.
And perhaps most disturbingly, in many places in the US, criminals do not have the right to vote, even after they are released. So much for "no taxation without representation".
However, most rights still have responsibilities. Freedom of speech, surely one of the most basic of basic rights in a democracy, does not include the freedom to engage in slander or libel. That's an associated responsibility.
For the record, I'm specifically talking about a general computer science education, with maths. Now read on...
No, it doesn't, but how can you make an informed choice between two path-finding algorithms? No, you don't have to have seen and analysed them both before, but you at least need to be able to do the analysis.
Again, sure. But consider the next generation of lock-free programming. To understand that, you need to be able to reason through cases, understand transactions, and have a pretty thorough knowledge of how the different levels computer cache (including store buffers) work.
What I'm saying is that a specific "games-only" web qualification is no replacement for a decent general computer science education. To be a modern software developer, be it in games or whatever, you need to know how to learn and adapt, far more than you need to know specific industry-specific factoids. And any company that values the latter over the former is not a company that you want to work for.
And if you want to implement AI? What if you're working for a company that isn't interested in last years' games, but next years', which might feature fluid simulation or something? And what about the next generation of multi-core CPUs where locking will be so expensive that you'll need to start using lock-free programming?
It might just be me, but I would not hire anyone who deliberately learned as little as possible, to do only just what was required to do last years' job.
The best favour you can do yourself is get a general education in computer science, with a thorough grounding in software engineering and pure/applied maths.
It's turtles all the way down, I tell you!
You're thinking like a slashdotter. Think like an open source businessperson for a moment, instead, and you'll see what's probably the most compelling reason, which is Checkbox completeness.
I can sell you a solution with the "yes, your Windows admins don't need to be retrained" checkbox checked. I can do that without paying the Redmond tax, and like all other open source, I can diagnose and fix problems by myself without being at the mercy of someone else's support cycle, assuming they want to fix your problem at all.
Did you forget to activate it? See http://reactosgenuineadvantage.com/ for details.
For those who don't want to read all the comments, here's the summary:
One example is skip lists, as already mentioned. Another common use (yes, I said common use; these things are all over software) is where you have data structures which logically need to be members of more than one set.
Looking at sched.h in my Linux 2.6 kernel source, for example, I see that a task_struct is threaded with no less than six doubly-linked and three singly-linked lists. That's an eleventy-linked data structure, for those keeping count, which makes your measly patented triply-linked list look positively sparse.