You really need to understand the java jit* that Suns java uses, because that 10 to 50 times slowdown only happens if you interprete the code instead of letting the jit handle it**. If you do enable the jit your 10 to 50 times will turn into something like 10% to 50% instead***.
*The jit will not compile the code the first time it reach it, because if the code is only executed once, interpreting it will be faster then compiling it and then running the compiled code. A good and fair test would be to implement 2 different sort algoritmems (qsort and mergesort for example) and then run them on a large dataset(Say 1GB integers). If you do this, please post the code and benchmark result.
**Which is only a problem in micro benchmarks, not in the real world.
***Based on all the java benchmarks I have seen. You can Google them your self. Start with Java Language shootout.
1: You are right. I forgot they made their own class file format.
Would buying a license to suns embedded java implementation get Google of the hook, or does the license require that you skip their entire stack with no modifications?
Google have implemented 1 and 2 according to the specs. But they have not implemented the entire java standard library because much of it(Such as awt and swing) would be an insane waste of resources to include in a phone.
But all that is needed for an Android phone to run standard java software is an implementation of the standard library included with jdk 1.6 but as I said nobody is insane enough to propose that. Not even Sun/Oracle is insane enough to propose that you should use the java standard edition on mobile phones.
I never managed to make the audio system in Windows XP behave as i would like. I have 2 sound output devices(Headphones and an usb headset) and I never managed to control which software did output to which sound device. And I also miss a Windows application to control the sound level on an "per application" level. And features such as redirect audio to other computers.
So part of the reason the linux audio system is still in a bit of a flux is that it is designed to do so much more then windows, thus making it much more difficult to design and implement. The other reason is that because there have newer been a single audio system/api different apps use different apis and this causes each sound system to add api emulation for the other sound systems which is quite a mess, and the emulation is not always that god.
Pulse audio is quite good until you use an old legacy app which only support oss. Then it will use the oss emulation api in pulseaudio-oss which atleast for me causes more then 1 second audio delay because the audio buffer is not flushed until it is full.
How about "You need to sell the company (Skype) because you need the money for other projects" and an IPO may be the best way to sell a company.
Or how about "We need to focus on our core business, and there is no synergy between an auction site and a ip phone company so we sell it. (Being owned by Ebay may make it difficult for Skype to do some thing they need to do, so selling it may be the best business thing to do.
But we are talking Skype a company with many non-paying customers and almost no income and revenue so I guess your option 2: "Let's dump it before other people realize how much overvalued Skype is" is the most likely answer.
I don't think he is hold to higher standards. If a paranoid office manager for example had changed all locks in the office*, and he was the only one with the new keys, he would have faced the same sense if he refused to give the keys when to his boss.
My responsibilities and duties as an IT worker end the moment I quit or someone fires me
They do with the exception that you have to deliver anything you have which belong to the company back to the company. So if you are fired you still have to hand back that company laptop, even if you currently store it at home. Passwords are not that different.
But as a competent sysadmin, I hope you don't admin any systems where you are the only one with the passwords. And I really hope there are systems where you store the config in a way where you are the only one with access. Remember: If you were paid to make it by the company it belong to the company and must be returned upon termination. This goes for both hardware, configuration files and other data you store which belong to the company.
*You may laugh, but something close to this did happen. But the guy did hand over the keys when the police came to get them and that was the end of that case.
Yes it does. It makes it very difficult to use hotkeys (Mostly for the non main window) because the main window keep getting focus when you use the mouse to work with it.
This really sucks if you use layers where I would so much love to let the layers window keep focus no matter what I do.
What I would really like is a "Don't ever give (keyboard) focus to this window and a "lock focus on current window until I explicit unlock the focus". But kde 4.4 really seems to miss thoose features, and so does gimp.
The way gimp does things may be cool if anyone would write a window manager with the needed features and include it with kde. (No I am not going away from kde, just to use gimp).
1: No absolutely not. My experience working for some rather large companies is that they are not really more effective then the government in general. One big problem the government have is that they can't hide their fault from the public the same way that a corporation can. And that the public care much more.
So when a large shipping company has to drop(And then redo all development) of their internal shipping system costing them >100 million dollors, all the publicity they get is an interesting article about prototype based development(Including a few things not to do).
When the government fuck up an computer project in the same size, they get multiple newspaper frontpages and often long and very public investigations about the cause. (Something good companies also do, but they keep it internal).
2: That is true. My rule of thumb is that very large organisations(Including the gov) have a rather large overhead but are still relative effective because their size allow them to access to the right(Or at least not total wrong) people to lead a project. Small companies can be really effective because of their much lower overhead, but they often relay on very few people doing the right thing to work. So sometimes they do really stupid things(Se: Thedailywtf.com) because they don't have access to any internal people who know anything about the subject.
Yes you can, but there seems to be a bit of difference between how Mysql and PostgreSQL define Repeatable read. According to postgresql (http://www.postgresql.org/docs/8.4/interactive/transaction-iso.html) the repeatable read does allow phantom read(That is: The same select executed in the same transaction can give 2 different results). According to mysql http://dev.mysql.com/doc/refman/5.0/en/set-transaction.html#isolevel_repeatable-read) this is not allowed in mysql.
If you want to avoid phantom-read in postgresql you need to specify the Serializable level, which seems to be same as repeatable read in mysql.
Oh and a funny fact: Postgresql don't have a specific repeatable read level, so if you request it, you will get Serializable but this is i guess subject to change.
But then we are talking about different solutions. The feature added to Postgresql is a shared nothing replication. A shared everything(It is just storage which is shared right?) will be able to handle synchronization much better.
But don't the RAC solution leave you with the storage system as a single point of failure then?
Quote: "If a Ford makes a truck, do they have to allow GM to have a compatible engine for it?"
Yes they do. If I buy the truck there is nothing that Ford can do if I install a GM engine in it.
And I search for "advertiser networks" in google I do in fact get advertising from a a company which describe them self as "Over 900 successful bloggers use BuySellAds.com to power their online ad sales. We help you sell ads better. We make your life easier."
But that is a bad example anyway, because there is an difference between a website you own, and a device you sell. I would not expect Apple to allow anyone to show advertising on their website or in Apples software, but neither would I expect Apple to have any control over what advertising 3 party software shows on devices once Apple have sold them(The device).
Does Oracle RAC really always ensure that if a query B is issued after* an other transaction A is comitted, then B will always see the result of A? Even if A and B are issued from different clients, and issued to different database servers?
*When looking at the wall clock. But it may in fact not be possible to say if B was really send before or after B comitet if the sender and reciveer of A and B are 4 different computers.
But why would users care? The only thing I as a user need from my bios is to boot my main operation system.
I always thought of the bios as a collection of hacks, which still manage to archive their job on each boot.
What purpose for example does the ability to install drivers serve? It might make it possible to use 3d graphics and 7 channel audio in the bios. But why?
And escaping data is not even enough. Look at this simple php code, and find the security bug (It is for mysql, but I guess the pg_query have the same problem)./* Asume an open connection to mysql exists This function allows any user who can submit to query any field in the database.
*/ function unsafe() {
$id=mysql_real_escape_string($_REQUEST['userId']);
$query="select name from user where id=$id";
$result=mysql_query($query);
while($mysql=mysql_fetch_row($result))
print_r($row); }
I really think the newspaper is right in this case and I can't belive why the blogger belive he is allowed to just put a copy of (large parts of) an article online on his own page. What he should do is to put a snippet(Normally the first paragraph or so) and then link to the newspaper. (Like google news does).
But when that is said, I still think that suing without even sending a warning is a bit aggresive.
Oh and the reason this kind of lawsuits are so seldom seen is that if the newspaper just send an takedown email, the article will normally be taken down, simply because the guy who put it up there normally know he has no right do so, so he will just take it down when asked.
Quote: e.g. Let's read the whole of RAM sequentially see how well your CPU cache performs.
That would be the absolutely best case usage for a cpu cache. Becasue when you read the first 32 bits, the cache will read the following 128/256 bits(Depending on cache line size/cpu/magic), which will then be ready to serve directly from cache. Much faster.
I saw a test where they added all the numbers in an 128MB array. They made 2 versions, one that iterated throu the array first to last, and one that did run last to first. And the one running last to first(That is: Running backwards in memory) were much(A factor 2 or 3 i think) slower.
Slashdot Mirror
You really need to understand the java jit* that Suns java uses, because that 10 to 50 times slowdown only happens if you interprete the code instead of letting the jit handle it**. If you do enable the jit your 10 to 50 times will turn into something like 10% to 50% instead***.
*The jit will not compile the code the first time it reach it, because if the code is only executed once, interpreting it will be faster then compiling it and then running the compiled code. A good and fair test would be to implement 2 different sort algoritmems (qsort and mergesort for example) and then run them on a large dataset(Say 1GB integers). If you do this, please post the code and benchmark result.
**Which is only a problem in micro benchmarks, not in the real world.
***Based on all the java benchmarks I have seen. You can Google them your self. Start with Java Language shootout.
Google made Wave* and GWT which are both quite innovative solutions.
*And then dropped it again.
1: You are right. I forgot they made their own class file format.
Would buying a license to suns embedded java implementation get Google of the hook, or does the license require that you skip their entire stack with no modifications?
The problem is that java really is:
1: A language
2: A vm
3: A standard class library
Google have implemented 1 and 2 according to the specs. But they have not implemented the entire java standard library because much of it(Such as awt and swing) would be an insane waste of resources to include in a phone.
But all that is needed for an Android phone to run standard java software is an implementation of the standard library included with jdk 1.6 but as I said nobody is insane enough to propose that. Not even Sun/Oracle is insane enough to propose that you should use the java standard edition on mobile phones.
How do I as a user choose that? Or do the app guys need to rewrite part of their app to use a new api?
I never managed to make the audio system in Windows XP behave as i would like. I have 2 sound output devices(Headphones and an usb headset) and I never managed to
control which software did output to which sound device. And I also miss a Windows application to control the sound level on an "per application" level. And features such as redirect audio to other computers.
So part of the reason the linux audio system is still in a bit of a flux is that it is designed to do so much more then windows, thus making it much more difficult to design and implement. The other reason is that because there have newer been a single audio system/api different apps use different apis and this causes each sound system to add api emulation for the other sound systems which is quite a mess, and the emulation is not always that god.
Pulse audio is quite good until you use an old legacy app which only support oss. Then it will use the oss emulation api in pulseaudio-oss which atleast for me causes more then 1 second audio delay because the audio buffer is not flushed until it is full.
How about "You need to sell the company (Skype) because you need the money for other projects" and an IPO may be the best way to sell a company.
Or how about "We need to focus on our core business, and there is no synergy between an auction site and a ip phone company so we sell it. (Being owned by Ebay may make it difficult for Skype to do some thing they need to do, so selling it may be the best business thing to do.
But we are talking Skype a company with many non-paying customers and almost no income and revenue so I guess your option 2: "Let's dump it before other people realize how much overvalued Skype is" is the most likely answer.
I don't think he is hold to higher standards. If a paranoid office manager for example had changed all locks in the office*, and he was the only one with the new keys, he would have faced the same sense if he refused to give the keys when to his boss.
My responsibilities and duties as an IT worker end the moment I quit or someone fires me
They do with the exception that you have to deliver anything you have which belong to the company back to the company. So if you are fired you still have to hand back that company laptop, even if you currently store it at home. Passwords are not that different.
But as a competent sysadmin, I hope you don't admin any systems where you are the only one with the passwords. And I really hope there are systems where you store the config in a way where you are the only one with access. Remember: If you were paid to make it by the company it belong to the company and must be returned upon termination. This goes for both hardware, configuration files and other data you store which belong to the company.
*You may laugh, but something close to this did happen. But the guy did hand over the keys when the police came to get them and that was the end of that case.
select $FIELDNAME,count(*) from $TABLENAME group by $FIELDNAME having count(*)>1
What is the developer supposed to use the remaining 50 seconds to?
Yes it does. It makes it very difficult to use hotkeys (Mostly for the non main window) because the main window keep getting focus when you use the mouse to work with it.
This really sucks if you use layers where I would so much love to let the layers window keep focus no matter what I do.
What I would really like is a "Don't ever give (keyboard) focus to this window and a "lock focus on current window until I explicit unlock the focus". But kde 4.4 really seems to miss thoose features, and so does gimp.
The way gimp does things may be cool if anyone would write a window manager with the needed features and include it with kde. (No I am not going away from kde, just to use gimp).
I think we just found the real reason Apple don't want iPhone to multitask
1: No absolutely not. My experience working for some rather large companies is that they are not really more effective then the government in general. One big problem the government have is that they can't hide their fault from the public the same way that a corporation can. And that the public care much more.
So when a large shipping company has to drop(And then redo all development) of their internal shipping system costing them >100 million dollors, all the publicity they get is an interesting article about prototype based development(Including a few things not to do).
When the government fuck up an computer project in the same size, they get multiple newspaper frontpages and often long and very public investigations about the cause. (Something good companies also do, but they keep it internal).
2: That is true. My rule of thumb is that very large organisations(Including the gov) have a rather large overhead but are still relative effective because their size allow them to access to the right(Or at least not total wrong) people to lead a project. Small companies can be really effective because of their much lower overhead, but they often relay on very few people doing the right thing to work. So sometimes they do really stupid things(Se: Thedailywtf.com) because they don't have access to any internal people who know anything about the subject.
I newer understod why InnoDB should allow filesystem snapshots. Is this documented anywore or is it just one of "It ought to work"?.
I mean InnoDB does NOT support a normal filesystem copy of the 3 database files as a way to make a backup.
Yes you can, but there seems to be a bit of difference between how Mysql and PostgreSQL define Repeatable read. According to postgresql (http://www.postgresql.org/docs/8.4/interactive/transaction-iso.html) the repeatable read does allow phantom read(That is: The same select executed in the same transaction can give 2 different results). According to mysql http://dev.mysql.com/doc/refman/5.0/en/set-transaction.html#isolevel_repeatable-read) this is not allowed in mysql.
If you want to avoid phantom-read in postgresql you need to specify the Serializable level, which seems to be same as repeatable read in mysql.
Oh and a funny fact: Postgresql don't have a specific repeatable read level, so if you request it, you will get Serializable but this is i guess subject to change.
But then we are talking about different solutions. The feature added to Postgresql is a shared nothing replication. A shared everything(It is just storage which is shared right?) will be able to handle synchronization much better.
But don't the RAC solution leave you with the storage system as a single point of failure then?
Quote:
"If a Ford makes a truck, do they have to allow GM to have a compatible engine for it?"
Yes they do. If I buy the truck there is nothing that Ford can do if I install a GM engine in it.
And I search for "advertiser networks" in google I do in fact get advertising from a a company which describe them self as
"Over 900 successful bloggers use BuySellAds.com to power their online ad sales. We help you sell ads better. We make your life easier."
But that is a bad example anyway, because there is an difference between a website you own, and a device you sell. I would not expect Apple to allow anyone to show advertising on their website or in Apples software, but neither would I expect Apple to have any control over what advertising 3 party software shows on devices once Apple have sold them(The device).
Does Oracle RAC really always ensure that if a query B is issued after* an other transaction A is comitted, then B will always see the result of A? Even if A and B are issued from different clients, and issued to different database servers?
*When looking at the wall clock. But it may in fact not be possible to say if B was really send before or after B comitet if the sender and reciveer of A and B are 4 different computers.
Is there any specific reason you did not just download the postgresql binary from their website, instead of building it yourself?
But why would users care? The only thing I as a user need from my bios is to boot my main operation system.
I always thought of the bios as a collection of hacks, which still manage to archive their job on each boot.
What purpose for example does the ability to install drivers serve? It might make it possible to use 3d graphics and 7 channel audio in the bios. But why?
The line
while($mysql=mysql_fetch_row($result))
should be
while($row=mysql_fetch_row($result))
Dam why don't mysql have an edit/append feature.
And escaping data is not even enough. Look at this simple php code, and find the security bug (It is for mysql, but I guess the pg_query have the same problem). /* Asume an open connection to mysql exists
This function allows any user who can submit to query any field in the database.
*/
function unsafe() {
$id=mysql_real_escape_string($_REQUEST['userId']);
$query="select name from user where id=$id";
$result=mysql_query($query);
while($mysql=mysql_fetch_row($result))
print_r($row);
}
Except that there is a "1 child" policy in most of China. So they should instead make porn legal, and hope people would have less real sex*.
*Based on the porn available in the western world, and the amount of sex people have, this does however not work :}
I really think the newspaper is right in this case and I can't belive why the blogger belive he is allowed to just put a copy of (large parts of) an article online on his own page. What he should do is to put a snippet(Normally the first paragraph or so) and then link to the newspaper. (Like google news does).
But when that is said, I still think that suing without even sending a warning is a bit aggresive.
Oh and the reason this kind of lawsuits are so seldom seen is that if the newspaper just send an takedown email, the article will normally be taken down, simply because the guy who put it up there normally know he has no right do so, so he will just take it down when asked.
Quote:
e.g. Let's read the whole of RAM sequentially see how well your CPU cache performs.
That would be the absolutely best case usage for a cpu cache. Becasue when you read the first 32 bits, the cache will read the following 128/256 bits(Depending on cache line size/cpu/magic), which will then be ready to serve directly from cache. Much faster.
I saw a test where they added all the numbers in an 128MB array. They made 2 versions, one that iterated throu the array first to last, and one that did run last to first. And the one running last to first(That is: Running backwards in memory) were much(A factor 2 or 3 i think) slower.
Or maybe i * detecter is just broken.
The difference beeing that Wal-Mart don't prevent other from selling things they don't sell.