When it comes down to it, languages are just wrappers around machine code. The wrappers are defined by the syntax of the programming language, and to a greater or lesser extent thats the only way in which they differ.
"Hence what, for lack of a better name, I'll call the Python paradox: if a company chooses to write its software in a comparatively esoteric language, they'll be able to hire better programmers, because they'll attract only those who cared enough to learn it."
And there in lies the greatest paradox. If companies employed people who knew python, everyone would learn it!!
What he's really advocating is use of unknown programming languages. Thats fine, but maitainability might become an issue simply because it is hard to find the people that code in it.
No-ones been hurt by sticking with the mainstream.
To me, thats a very strange statement. The standard library of Java is, IMHO, one of the strongest around.
If you combine the java Collections with the Jakarta Collections library then you have an almost unbeatable combination (beaten perhaps only by Lisp and its treatement of collections, lists etc).
I admit that there are a myriad of redundant and (mostly) confusing and unnecessary standards. But you shouldn't complain that so many XML parsing toolkits exist, that gives you the freedom to choose the one thats right for a given application.
And after all, isnt that what this is all about? Categorising one programmer as better than another because of their programming language is like saying that surgeons are better than barbarians. Both use blades, but you dont want to perform heart operations with a longsword!
No, my point is that this Isnt a virus, and its not anything special.
If using windows, go to a command console and type format c:\. Linux? rm -rf/.
Is what I did there a virus? I exectuded code of my own volition. We've had the power to destroy computers through our own malice/stupidity for eons, this is no different. Why do I care that I could damage my phone by running code, because thats what this is. Its stupidity on the part of the user, not malignant self propogating code.
This is more a user intelligence program than a true threat to the symbian 60 series. If it propogated to all the numbers in a phone book (via SMS for example) then it would be something worth worrying about.
Stored Procedures. They arent functions. Functions are different. Functions should be called inline, from within SQL statements. There should be a difference.
When looking towards migrating to an OSS database from MS SQL Server I looked into how easy it was to use the postgres stored procedure/functions/things. I couldn't find any equivalent of returing a resultset. In MS SQL you use:
CREATE PROCEDURE name AS select * from test
The only way that I could find to do that in postgres was:
Return a cursor reference, in which case throuch JDBC you need to obtain this horrid postgres specific class from the driver.
Return a set of results (or was it records, my mind fails me), but I couldn't get this to work
Return a set of created types, but this doesnt work well for dynamic queries and makes maintainance a headache - you'd end up defining hundreds of types!
So it seems that we're stuck with MS SQL server.
If someone DOES know how to do this is a non-evil manner, please tell me! Were planning on doing an upgrade, and I'd rather not have to fork out the money for SQL Server licences etc...
Yep, theres a lot to be said for this.
In fact, now I think about it, Asimov touch upon this in the Foundation series. The only way psychohistory works is that people don't know that it exists. If everyone could read the future, by definition all of their predictions would be wrong.
With regards to point 4) She would need to sell it to someone called Katie, otherwise Penguin may have a legitimate claim that said person was squatting.
Note the may. Lawyers can argue that they have more right to it than the owner does, or that the owner is stealing their publicity for their own purposes.
Hell, with 10 megs of patches going into the 2.6 kernel tree per month, Its a wonder there aren't more nasties in there. Its gonna take some review body to QA that volume of code.
Thats exactly my point. I can't write Java and have it guaranteed to compile and execute the same way on any of those JVM's that you point out. Hence they arent true JVM's.
Oh yes you can make a little hobby project that compiles java into native code but without the accreditation of passing the TCK and being certified as a true JVM, businesses and others wont use it!
There is an interesting piece on Anne Thomas Manes' weblog, which I shall copy below:
I was talking to Toufic Boubez yesterday, the CTO of Layer 7. Layer 7 provides a Web services policy and security management appliance, and they use a hardened Linux operating system. I asked him why he doesn't use BSD. And he said
Because there's no decent JVM for BSD!
If Java were truly open source, then this wouldn't be a problem. You could just take the source and compile your own JVM for BSD. But no. If you want to create a JVM for BSD, you have to be a licensee of Java (big bucks!). None of the current licensees of Java has an incentive to develop a JVM for BSD, hence no JVM
The reason for this is that Sun charges very high license fees for people wanting to develop JVMs. This is why there is no JVM for FreeBSD yet (or so I have been told).
Anything that carries the Java name must be java compliant. To be java compliant it must have passed the Java TCK published by Sun. The TCK is only available to licensees.
I don't believe that open sourcing Java is a good idea - if anyone could fork the codebase in any way they want versioning becomes an issue and you potentially loose cross JVM compatability, not just cross platform.
They are? If I hear of root exploit, I take all harmed services down immediately, and patch one by one. I also give calls to the companies I work with. They agree that having a little bit of downtime is well worth the risk of not being auto-hacked.
You miss the point. You are an educated and clued up person with respect to operating systems in general. As more and more people are migrated to linux so the quality of the userbase goes down. People will still open stupid attachments and find a way of getting round security even on Linux.
Any company that relies on its employees to keep current with AntiVirus updates and patches is a fool. Hence such things MUST be an administration issue. The platform does not matter, patches etc still need to be applied.
The bottom line? A computer is as secure as the person thats using/administering it, not matter what the O/S.
I am NOT trying to spread Microsoft FUD; I am a linux advocate. I do believe, however, that linux advocates are going to have a shock when linux does reach 50% market penetration - yes viruses will be written, exploits will be used, people will install dumb shit on their computers and open the door for spyware. Linux is, after all, just another operating system. Its just as good as the person that secures it
A hell of a lot of people dont patch systems because doing so will break compatability between products. That mindset will still continue when working with linux and so systems will still go unpatched (along with other reasons such as under-funding, indifference, lack of time etc). I know this because my home gentoo box has a couple of vulnerabilities I need to sort out. But I havent.
I'm the kind of person that linux is going to have to deal with. I browse the web from root! Try explaining to your boss that, no, infact he cant do X because it violates the security policy, or that he has to change users to install stuff.
-O3 won't allways help you for optimisations. -03 generates huge binaries. This means that, often, the relevant sections of the program cannot be kept in the processor's cache and thus must be shifted out to ram resulting in slower execution.
-Os, which seeks to minimise the size of generated files, often manages to keep more of the application in cache and this leads to an increase in execution speed.
-O3 is really useful for people with 1 or 2mb of on-die cache processors. Who knows, in the future these may become abundant. It's swings and roundabouts; the benefits of one -O over another have been beaten to death.
As an aside issue, the actual machine code generated by the JIT compiler isnt that great. Whats brilliant is that the compiler profiles in realtime, optimising the code for the input by heavily optimising hot branches and caching them. This often generates a 30% odd increase in performance.
As a Windows user I don't care why Windows is faster for gui stuff, I just care that it is.
If loading openoffice into memory is what it takes then load openoffice into memory! (Although Im sure that this would add to the allready large system startup time that's also a problem with linux.)
I run a dual boot Gentoo system and the time to boot up and get to OpenOffice is an order of magnitude greater than the time to boot up XP and get into Word.
What you need to learn is that spelling & grammar != intelligence. I know several people who are smarter then me, but who can't spell worth crap. At least one is dyslexic, the others may just have actual important things on there minds.
What good is intelligence if you can't communicate your ideas properly?
Did the people who coded MySQL also design the C programming language? Or in your analogy, did the guy who invented the drivetrain on the Buick also discover how to mine iron ore, smelt it down, then cast it?
Innovation _allways_ builds upon something else.
Slashdot Mirror
You wont bare it? Thank god for that!
The court told them where to go to make sure that the court decision was overruled!
Completely off topic, I admit, but I've seen your sig before and cant work out what it does. What does it print out?
I really want the mod to explain why he decided that the parent was a Troll.
Now this post? Flaimbait probably :p
When it comes down to it, languages are just wrappers around machine code. The wrappers are defined by the syntax of the programming language, and to a greater or lesser extent thats the only way in which they differ.
And there in lies the greatest paradox. If companies employed people who knew python, everyone would learn it!!
What he's really advocating is use of unknown programming languages. Thats fine, but maitainability might become an issue simply because it is hard to find the people that code in it.
No-ones been hurt by sticking with the mainstream.
If you combine the java Collections with the Jakarta Collections library then you have an almost unbeatable combination (beaten perhaps only by Lisp and its treatement of collections, lists etc).
I admit that there are a myriad of redundant and (mostly) confusing and unnecessary standards. But you shouldn't complain that so many XML parsing toolkits exist, that gives you the freedom to choose the one thats right for a given application.
And after all, isnt that what this is all about? Categorising one programmer as better than another because of their programming language is like saying that surgeons are better than barbarians. Both use blades, but you dont want to perform heart operations with a longsword!
If using windows, go to a command console and type format c:\. Linux? rm -rf /.
Is what I did there a virus? I exectuded code of my own volition. We've had the power to destroy computers through our own malice/stupidity for eons, this is no different. Why do I care that I could damage my phone by running code, because thats what this is. Its stupidity on the part of the user, not malignant self propogating code.
This is more a user intelligence program than a true threat to the symbian 60 series. If it propogated to all the numbers in a phone book (via SMS for example) then it would be something worth worrying about.
Thanks, thats definitely something to think about when it comes to the upgrade. I did read that techdocs article, but couldnt get it to work.
Suppose that I said
Can I pass parameters to views? Thats not a rhetorical question, I'm really not sure!!
Stored Procedures. They arent functions. Functions are different. Functions should be called inline, from within SQL statements. There should be a difference.
When looking towards migrating to an OSS database from MS SQL Server I looked into how easy it was to use the postgres stored procedure/functions/things. I couldn't find any equivalent of returing a resultset. In MS SQL you use:
The only way that I could find to do that in postgres was:
So it seems that we're stuck with MS SQL server.
If someone DOES know how to do this is a non-evil manner, please tell me! Were planning on doing an upgrade, and I'd rather not have to fork out the money for SQL Server licences etc...
Yep, theres a lot to be said for this.
In fact, now I think about it, Asimov touch upon this in the Foundation series. The only way psychohistory works is that people don't know that it exists. If everyone could read the future, by definition all of their predictions would be wrong.
Note the may. Lawyers can argue that they have more right to it than the owner does, or that the owner is stealing their publicity for their own purposes.
Hell, with 10 megs of patches going into the 2.6 kernel tree per month, Its a wonder there aren't more nasties in there. Its gonna take some review body to QA that volume of code.
I might be tempted to use it if SOMEONE could give me a reason why it's better than Ant and Forrest in combination.
Oh yes you can make a little hobby project that compiles java into native code but without the accreditation of passing the TCK and being certified as a true JVM, businesses and others wont use it!
There is an interesting piece on Anne Thomas Manes' weblog, which I shall copy below:
The key here is decent.
Anything that carries the Java name must be java compliant. To be java compliant it must have passed the Java TCK published by Sun. The TCK is only available to licensees.
I don't believe that open sourcing Java is a good idea - if anyone could fork the codebase in any way they want versioning becomes an issue and you potentially loose cross JVM compatability, not just cross platform.
You miss the point. You are an educated and clued up person with respect to operating systems in general. As more and more people are migrated to linux so the quality of the userbase goes down. People will still open stupid attachments and find a way of getting round security even on Linux.
Any company that relies on its employees to keep current with AntiVirus updates and patches is a fool. Hence such things MUST be an administration issue. The platform does not matter, patches etc still need to be applied.
The bottom line? A computer is as secure as the person thats using/administering it, not matter what the O/S.
I am NOT trying to spread Microsoft FUD; I am a linux advocate. I do believe, however, that linux advocates are going to have a shock when linux does reach 50% market penetration - yes viruses will be written, exploits will be used, people will install dumb shit on their computers and open the door for spyware. Linux is, after all, just another operating system. Its just as good as the person that secures it
A hell of a lot of people dont patch systems because doing so will break compatability between products. That mindset will still continue when working with linux and so systems will still go unpatched (along with other reasons such as under-funding, indifference, lack of time etc). I know this because my home gentoo box has a couple of vulnerabilities I need to sort out. But I havent.
I'm the kind of person that linux is going to have to deal with. I browse the web from root! Try explaining to your boss that, no, infact he cant do X because it violates the security policy, or that he has to change users to install stuff.
-Os, which seeks to minimise the size of generated files, often manages to keep more of the application in cache and this leads to an increase in execution speed.
-O3 is really useful for people with 1 or 2mb of on-die cache processors. Who knows, in the future these may become abundant. It's swings and roundabouts; the benefits of one -O over another have been beaten to death.
As an aside issue, the actual machine code generated by the JIT compiler isnt that great. Whats brilliant is that the compiler profiles in realtime, optimising the code for the input by heavily optimising hot branches and caching them. This often generates a 30% odd increase in performance.
If loading openoffice into memory is what it takes then load openoffice into memory! (Although Im sure that this would add to the allready large system startup time that's also a problem with linux.)
I run a dual boot Gentoo system and the time to boot up and get to OpenOffice is an order of magnitude greater than the time to boot up XP and get into Word.
Thr trouble with verbal cotnracts is that they aren't worth the paper they're written on.
Did the people who coded MySQL also design the C programming language? Or in your analogy, did the guy who invented the drivetrain on the Buick also discover how to mine iron ore, smelt it down, then cast it?
Innovation _allways_ builds upon something else.