Sure. As the old saying goes: "Russia has only two friends: Russian army and Russian navy".
Nobody is going to believe in "international friendship" anymore. Everybody has their own interests, and right now Russia and Europe just have some common interests. That's OK.
Some people just don't understand how is it possible to work together with Russian without becoming a commie.
Let's see: 1) Cyanides (a glass of water with 1:4200 mix and you're as good as dead) 2) Dioxins (AFAIR, it causes cancer in 1:1000000000 ratios) 3) Good old benzene (carcinogen, 1:4200 is FAR above the safe threshold) 4) Thiols (1:4200 will make water STINK so much you'll faint)...
I generally use autogenerated schemas (with manually-added indexes and advanced integrity checks) when I'm using Hibernate. I don't remember the time when Hibernate offered bad schema. Usually mapping from object structure to SQL DDL is very simple (you map objects as tables, one-to-many relations as backpointers, etc.).
Decent ORMs do nothing but map object operations into SQL statements. SQL from an ORM tool is not going to magically work more faster or slower than a hand-written one.
Again, decent DB schemas (i.e. fairly normalized ones) map nicely into object models (hell, ER-diagrams used to model relational tables map directly into object diagrams). It's the databases with weird tables without PKs and strange stored procs which do not map well.
I've worked with Hibernate application handling OLAP operations on 10 terabytes of data without any problems.
Sure. But very often Mac OS X/Linux behave different from Windows for no good reason. We had numerous problems with focus handling, controls being one pixel off grid, etc.
It's easier with QT/GTK because they draw widgets themselves, so a lot of problem go away. If only GTK was not an ugly mess on Windows and Mac OS...
Nope. EACH developer who needs to compile sources which use QT should have a developer license (that's what our legal department said after talking with Trolltech). In practice, it's easier to buy license for every developer.
You can have your code under BSD license, because it's less restricting than GPL. But if anyone tries to use your application as a part of a commercial closed-source project, then they will be violating _GPL_ license of QT. Which, sort of, defeats the whole purpose of BSD license...
You can have QT in BSDs without GPLing the whole thing because of the 'aggregation' clause in GPL.
It's true that complexity of any pseudorandom generator is finite.
However, you can't readily tell it from its output. It might well be possible, that the Universe uses a simple formula to determine the outcome of quantum events.
In any case, we can just add hardware-based random number generator to classical computer simulating quantum events.
How neurons talk to each other is irrelevant - you can't use classical channels to connect quantum computers. And I can't imagine any quantum channel working across multiple cell walls.
QCs (if they are present inside neurons) are probably advantageous, but it doesn't mean that they will be the fundamental barrier for creating 'artificial brain'. However, it may be practically impossible to build artificial brain without some sort of QC accelerator (i.e. like videocards - you can build games which use only CPU, but it would be prohibitively slow in practice).
What is 'really random number'? Can you define it?
We can have pseudorandom numbers behaving exactly like real 'random' numbers (though it requires unlimited resources as you get more and more numbers). I.e. you won't be able to distinguish pseudorandom generator and real random source.
Nope, Penrose's quantum computers work on a sub-cellular level.
Thinking about branches requires work of more than one neuron, so quantum computers are not really significant here (they can't talk to each other through cell walls).
Penrose was probably smocking crack when he wrote this book.
You definitely don't NEED quantum computers to reproduce intelligence. That's because IF cells contain quantum computers, then they must work in cycles: load initial data, process it, read data. Reading computation results stops quantum computer (collapses it to one state). Even Penrose admits that quantum computers can't work more than a fraction of second in a living cell.
Quantum computers can be simulated by classical computers (they're computationally equivalent), so quantum computers are not NEEDED to simulate human mind.
However, quantum computers might make good accelerators for neural processing (there are several publications on this).
Slashdot Mirror
Sure. As the old saying goes: "Russia has only two friends: Russian army and Russian navy".
Nobody is going to believe in "international friendship" anymore. Everybody has their own interests, and right now Russia and Europe just have some common interests. That's OK.
Some people just don't understand how is it possible to work together with Russian without becoming a commie.
It's true - a lot of nasty things do not mix well with water. They float on water surface (hint: oil), suffocate fish and pollute shores.
Let's see: ...
1) Cyanides (a glass of water with 1:4200 mix and you're as good as dead)
2) Dioxins (AFAIR, it causes cancer in 1:1000000000 ratios)
3) Good old benzene (carcinogen, 1:4200 is FAR above the safe threshold)
4) Thiols (1:4200 will make water STINK so much you'll faint)
No, you can't.
TiVO patented their hardware and software, so you'd have to license it.
In all ORMs that I know, you can use plain SQL for complex queries. And it's usually much easier to use plain SQL than ORM's built-in query language.
But in my experience, 90% of queries are simple "SELECT blah.blah,... FROM blah WHERE blah.blah=?" and ORMs allow to forget about this simple CRUD.
You can have equally stupid example with SQL:
RecordSet r=db.query("SELECT * FROM product");
while(r.hasMoreRows())
{
Row row=r.nextRow();
if (row.get(0)<2.99)
myResults.push(row[0],row[1],row[2]);
}
I've seen enough of such WTFs, usually in PHP programs.
I generally use autogenerated schemas (with manually-added indexes and advanced integrity checks) when I'm using Hibernate. I don't remember the time when Hibernate offered bad schema. Usually mapping from object structure to SQL DDL is very simple (you map objects as tables, one-to-many relations as backpointers, etc.).
Your statement is just plain stupid.
Decent ORMs do nothing but map object operations into SQL statements. SQL from an ORM tool is not going to magically work more faster or slower than a hand-written one.
Again, decent DB schemas (i.e. fairly normalized ones) map nicely into object models (hell, ER-diagrams used to model relational tables map directly into object diagrams). It's the databases with weird tables without PKs and strange stored procs which do not map well.
I've worked with Hibernate application handling OLAP operations on 10 terabytes of data without any problems.
It's a knock-off from phage therapy ( http://en.wikipedia.org/wiki/Phage_therapy ) which is about 70 years old.
See http://kerneltrap.org/node/553/2131 for explanation. In short, Linus has good reasons to use goto.
Sure. But very often Mac OS X/Linux behave different from Windows for no good reason. We had numerous problems with focus handling, controls being one pixel off grid, etc.
It's easier with QT/GTK because they draw widgets themselves, so a lot of problem go away. If only GTK was not an ugly mess on Windows and Mac OS...
GTK looks like and behaves like s**t on Windows. We've tried to use it and had a LOT of problem.
The only other solution is wxWidgets, but it also has big problems with applications behaving differently on different platforms.
Nope. EACH developer who needs to compile sources which use QT should have a developer license (that's what our legal department said after talking with Trolltech). In practice, it's easier to buy license for every developer.
BTW, I was wrong with the price. It's $6600 per developer for three-platform desktop edition - http://trolltech.com/products/qt/licenses/pricing
No, you can't (effectively).
You can have your code under BSD license, because it's less restricting than GPL. But if anyone tries to use your application as a part of a commercial closed-source project, then they will be violating _GPL_ license of QT. Which, sort of, defeats the whole purpose of BSD license...
You can have QT in BSDs without GPLing the whole thing because of the 'aggregation' clause in GPL.
QT costs A LOT of money (about $3000 per developer, AFAIR). _AND_ you can't legally use KDE's functions, because KDE is GPLed.
So, GTK is the only viable alternative (and guess what, most commercial Linux apps use GTK).
Nope, it's all about how to AVOID ironing. Nerdy enough for me :)
Just tried to pay using my Visa. It works.
Actually, it seems that http://www.mp3sparks.com/ already offers all the music from AoMP3.
So I guess AoMP3 has already reincarnated.
3...2...1
GO!
It's true that complexity of any pseudorandom generator is finite.
However, you can't readily tell it from its output. It might well be possible, that the Universe uses a simple formula to determine the outcome of quantum events.
In any case, we can just add hardware-based random number generator to classical computer simulating quantum events.
How neurons talk to each other is irrelevant - you can't use classical channels to connect quantum computers. And I can't imagine any quantum channel working across multiple cell walls.
QCs (if they are present inside neurons) are probably advantageous, but it doesn't mean that they will be the fundamental barrier for creating 'artificial brain'. However, it may be practically impossible to build artificial brain without some sort of QC accelerator (i.e. like videocards - you can build games which use only CPU, but it would be prohibitively slow in practice).
What is 'really random number'? Can you define it?
We can have pseudorandom numbers behaving exactly like real 'random' numbers (though it requires unlimited resources as you get more and more numbers). I.e. you won't be able to distinguish pseudorandom generator and real random source.
Nope, Penrose's quantum computers work on a sub-cellular level.
Thinking about branches requires work of more than one neuron, so quantum computers are not really significant here (they can't talk to each other through cell walls).
AFAIR my university's course, Turing machine with random number generator is computationally equivalent to a common Turing machine.
Penrose was probably smocking crack when he wrote this book.
You definitely don't NEED quantum computers to reproduce intelligence. That's because IF cells contain quantum computers, then they must work in cycles: load initial data, process it, read data. Reading computation results stops quantum computer (collapses it to one state). Even Penrose admits that quantum computers can't work more than a fraction of second in a living cell.
Quantum computers can be simulated by classical computers (they're computationally equivalent), so quantum computers are not NEEDED to simulate human mind.
However, quantum computers might make good accelerators for neural processing (there are several publications on this).