True enough. Many laptops will have a firewire port that can directly address memory through DMA. I don't know how many hardware vendors are effected by firewire connections that can basically read all of the memory, but I think there might be quite a few around. Most people won't even know such holes exist, and I don't even know how widespread this problem is.
Most of the time these functions hook in on the Windows logon process. Of course you may still have an administrator with a pretty strong password to log in to your system if you are locked out.
Most of the times these functions exist to make it easier to logon/logoff. You'll do it sooner that way. I have to type my password many, many times a day because I am supposed to lock my machine whenever I am not around. After a while you stop doing so when you are going for the coffee machine (5m) then the toilet (10m), then the neighboring lab etc.
It's more secure than leaving your machine open, that's what's count. The other security is/should be build in by Windows (e.g. encryption of the users home folder for both the admin as well as the facial recognition.
One problem with this scheme is that in comparison with password security, it is not easy to derive a unique "secret" value from the input of a picture. There is some work done on this for fingerprint biology, but even that is still in the research stage (although it is some time ago that I've looked into that). So you basically rely on a compare with the data already present in the machine, and you cannot directly derive secret keys.
The other advantage of speech recognition software is that it might notice the difference between "there" and "their". On the other hand, it might not recognize the word "bizarro".
It's more tiresome that there is only a single thread updating the GUI & renderer. This is hell when e.g. Adobe installs itself as a plugin, especially since it halts it's own GUI process each and every time it has to do something.
Plugins should really run in different threads independent on the GUI thread doing the updating (different processes would make communication more difficult, but might be a good second choice), since the plugins would not be able to bring the browser down to its knees.
Each page and mayor subitem should really have its own rendering thread, then these stupid freezes would come to an end. They could just give their results to the main thread doing the (user visible) updating.
I really hate it when I download a viewer or player and it will default to all file formats, unless it is easy to disable that functionality *and* it is easy to reverse the operation. This allows me to test-drive different players. If this kind of functionality is not included, it won't take the first hurdle and I will tell everyone *not* to use it.
Most people that download firefox will probably want it as their default browser, it is only a single checkbox and easy to reverse. So firefox takes the hurdle with ease.
Slightly OT: The most obnoxious thing I've seen with default browsers is Microsoft/others *knowingly* using the default browser while a site is only created with IE in mind. This is very dirty warfare. Fortunately, they seem to have refrained from using this tactic lately (IE 7 would probably crash on a IE 6 only site anyway:).
Ah, my options were still set to "code". Stupid option anyway, if I want code included, it's in *addition* to normal text. I've not seen a single comment entirely consisting of code. See experts exchange for a better idea.
"Why is it that web designers and developers - and I'm guilty of this too - almost always knowingly use a browser that most of their users won't?"
Because it is (more) standards compliant. Because it is easier to code for. Because their are more tools available for it. Because you can have different versions of it on a system. Because they don't confuse the certificate store of the browser with that of the system. Because the other browser is simply more usable. Because they don't want to see all those adds. Because they want to have IE6 installed (as most of their unknowing users do) but want to browse without getting a trojan. Need more?
Note that this is for personal and professional browsing of internet standards and the like. It's not said that they don't test their pages using IE. You can be pretty sure that they do, I hardly ever see a page that renders badly on IE, even though their compliance to web standards sucks (or sucked) badly.
As for the choice of the specific non-proportional font please blame the SlashDot site. As for using a non-proportional font while discussing a CLI, well, this IS slashdot. Hand in your geek license *now*!
That's why you need a toolkit that directly connects to the language and knows how to convert from one to the other (keeping in line with the code conventions for that language). For Java there are a few Eclipse plugins that manage to do this.
Not that I like UML for designing my application, but that's another matter. Having a class design be generated automatically after implementation can be a huge win (if the design is not to complicated I will do it inside my head instead of using an overcomplicated square/circle drawing program, thank you).
But seriously, if your tool chain cannot handle re-factoring, you've got yourself a big problem.
Yes, and it would be interesting to create a time-line with points where a new framework was created and for which reason. Most of the time a framework is created because something is very hard to do using the defined J2EE process. J2EE is updated continuously however, integrating e.g. hybernate functionality and other framework capabilities. It might not hurt just to stick to J2EE because you could do most things using J2EE.
But the explosion of frameworks has happened, and I must admit that it is not easy to keep up with it. At least not when you spend most of your time doing J2SE + cryptography as I'm doing now.
Yup, they are behind with their chipset offering. I do expect that they right this - and hopefully for them pretty soon. Otherwise Intel has a very interesting chip with a chipset that's just a stripped down desktop part, behaving pretty badly.
For an average consumer, the chipset is at least as important as the CPU, so I never understand how they can come up with such an interesting chip design, just to pair it off with a dinosaur.
"Does this apply to a particular CPU/MMU compiler combination or is it generic across all systems? Is it technically possible to design a system that is immune to buffer overruns or, by default, fails safe, as in not allowing any old code to walk all over the address space."
Yes, it's called managed code (Java/.NET) and yes, you can even design hardware that runs byte code. It will slightly hamper performance, but it has its advantages. Of course, the way it is currently done is to implement the JVM in software. That's ok though, you have such a small target running unsafe code that the number of buffer overruns is insignificant.
When there is a problem, an exception is raised. But an exception is a basic component in the byte code and it just crashes that part of the system at worst. Obviously that does not mean you cannot create mistakes when using managed code, but they tend not to spread as far.
Together with a good messaging system and/or immutable objects, you can create a heck of a safe system.
Yes, well, this happens to any language + set of API's. I won't defend Java here, there are way too many web-frameworks and data persistence frameworks to choose from. This happens when people find a gap in usability and try to fill it, each in a different way with slightly different benefits.
The good thing about Java is that you could just stick to J2SE + J2EE and be done with it. You should be able to do most things without the problem, but you would miss out on the benefits the other frameworks may offer. Most of the time you will just have to wait a bit and the good functionality gets incorporated in the API's and language features, and - most of the time - in a very nice way.
Both of your comments show how strong the language is. And why I would never *ever* like to read a few thousand lines of your code. YUCK.
Take off new line *especially*? WTF?
l = [int(x.strip()) for x in l.strip().split(',')]
Are you really trying to say I should prefer this to the methods available in Java? I (*and* you) would easily understand a piece of code doing the same thing in Java.
I've had the same issues with C/C++ and a lot of other languages though. Most Java programs run fine on the latest JDK, and some run up to 20/30% faster on JDK 1.6. There are a few frameworks that try to cope with this problem, such as the OSGi framework for Java and assemblies for.NET. But it is a complex problem and I don't think *any* language has solved this problem completely. It might not even be completely solvable, come to think of it.
My code compiled against 1.2 still runs fine in JDK 1.6 though. So I cannot call that horrible or anything.
Slashdot Mirror
In other words: bill, I'm still getting spam in my mailbox. What gives?
I don't know if he's a window user, but he's certainly a backslashing sob!
Yeah, I was wondering what Gaudi was doing so far north myself :)
True enough. Many laptops will have a firewire port that can directly address memory through DMA. I don't know how many hardware vendors are effected by firewire connections that can basically read all of the memory, but I think there might be quite a few around. Most people won't even know such holes exist, and I don't even know how widespread this problem is.
Most of the time these functions hook in on the Windows logon process. Of course you may still have an administrator with a pretty strong password to log in to your system if you are locked out.
Most of the times these functions exist to make it easier to logon/logoff. You'll do it sooner that way. I have to type my password many, many times a day because I am supposed to lock my machine whenever I am not around. After a while you stop doing so when you are going for the coffee machine (5m) then the toilet (10m), then the neighboring lab etc.
It's more secure than leaving your machine open, that's what's count. The other security is/should be build in by Windows (e.g. encryption of the users home folder for both the admin as well as the facial recognition.
One problem with this scheme is that in comparison with password security, it is not easy to derive a unique "secret" value from the input of a picture. There is some work done on this for fingerprint biology, but even that is still in the research stage (although it is some time ago that I've looked into that). So you basically rely on a compare with the data already present in the machine, and you cannot directly derive secret keys.
The other advantage of speech recognition software is that it might notice the difference between "there" and "their". On the other hand, it might not recognize the word "bizarro".
In many ways it would make an interesting read.
It's more tiresome that there is only a single thread updating the GUI & renderer. This is hell when e.g. Adobe installs itself as a plugin, especially since it halts it's own GUI process each and every time it has to do something.
Plugins should really run in different threads independent on the GUI thread doing the updating (different processes would make communication more difficult, but might be a good second choice), since the plugins would not be able to bring the browser down to its knees.
Each page and mayor subitem should really have its own rendering thread, then these stupid freezes would come to an end. They could just give their results to the main thread doing the (user visible) updating.
I really hate it when I download a viewer or player and it will default to all file formats, unless it is easy to disable that functionality *and* it is easy to reverse the operation. This allows me to test-drive different players. If this kind of functionality is not included, it won't take the first hurdle and I will tell everyone *not* to use it.
:).
Most people that download firefox will probably want it as their default browser, it is only a single checkbox and easy to reverse. So firefox takes the hurdle with ease.
Slightly OT: The most obnoxious thing I've seen with default browsers is Microsoft/others *knowingly* using the default browser while a site is only created with IE in mind. This is very dirty warfare. Fortunately, they seem to have refrained from using this tactic lately (IE 7 would probably crash on a IE 6 only site anyway
Ah, my options were still set to "code". Stupid option anyway, if I want code included, it's in *addition* to normal text. I've not seen a single comment entirely consisting of code. See experts exchange for a better idea.
"Why is it that web designers and developers - and I'm guilty of this too - almost always knowingly use a browser that most of their users won't?"
Because it is (more) standards compliant. Because it is easier to code for. Because their are more tools available for it. Because you can have different versions of it on a system. Because they don't confuse the certificate store of the browser with that of the system. Because the other browser is simply more usable. Because they don't want to see all those adds. Because they want to have IE6 installed (as most of their unknowing users do) but want to browse without getting a trojan. Need more?
Note that this is for personal and professional browsing of internet standards and the like. It's not said that they don't test their pages using IE. You can be pretty sure that they do, I hardly ever see a page that renders badly on IE, even though their compliance to web standards sucks (or sucked) badly.
As for the choice of the specific non-proportional font please blame the SlashDot site. As for using a non-proportional font while discussing a CLI, well, this IS slashdot. Hand in your geek license *now*!
That bowing seems to be an empty statement.
Try laughing until you cry :)
That's why you need a toolkit that directly connects to the language and knows how to convert from one to the other (keeping in line with the code conventions for that language). For Java there are a few Eclipse plugins that manage to do this.
Not that I like UML for designing my application, but that's another matter. Having a class design be generated automatically after implementation can be a huge win (if the design is not to complicated I will do it inside my head instead of using an overcomplicated square/circle drawing program, thank you).
But seriously, if your tool chain cannot handle re-factoring, you've got yourself a big problem.
Yes, and it would be interesting to create a time-line with points where a new framework was created and for which reason. Most of the time a framework is created because something is very hard to do using the defined J2EE process. J2EE is updated continuously however, integrating e.g. hybernate functionality and other framework capabilities. It might not hurt just to stick to J2EE because you could do most things using J2EE.
But the explosion of frameworks has happened, and I must admit that it is not easy to keep up with it. At least not when you spend most of your time doing J2SE + cryptography as I'm doing now.
Yup, they are behind with their chipset offering. I do expect that they right this - and hopefully for them pretty soon. Otherwise Intel has a very interesting chip with a chipset that's just a stripped down desktop part, behaving pretty badly.
For an average consumer, the chipset is at least as important as the CPU, so I never understand how they can come up with such an interesting chip design, just to pair it off with a dinosaur.
Anyway, afaik, they are working on it.
"Does this apply to a particular CPU/MMU compiler combination or is it generic across all systems? Is it technically possible to design a system that is immune to buffer overruns or, by default, fails safe, as in not allowing any old code to walk all over the address space."
Yes, it's called managed code (Java/.NET) and yes, you can even design hardware that runs byte code. It will slightly hamper performance, but it has its advantages. Of course, the way it is currently done is to implement the JVM in software. That's ok though, you have such a small target running unsafe code that the number of buffer overruns is insignificant.
When there is a problem, an exception is raised. But an exception is a basic component in the byte code and it just crashes that part of the system at worst. Obviously that does not mean you cannot create mistakes when using managed code, but they tend not to spread as far.
Together with a good messaging system and/or immutable objects, you can create a heck of a safe system.
Yes, well, this happens to any language + set of API's. I won't defend Java here, there are way too many web-frameworks and data persistence frameworks to choose from. This happens when people find a gap in usability and try to fill it, each in a different way with slightly different benefits.
The good thing about Java is that you could just stick to J2SE + J2EE and be done with it. You should be able to do most things without the problem, but you would miss out on the benefits the other frameworks may offer. Most of the time you will just have to wait a bit and the good functionality gets incorporated in the API's and language features, and - most of the time - in a very nice way.
"...you pass this indentation test..."
There, fixed it for you.
(did I pass?)
Except when the variable lives longer than the for loop, of course :)
Come back when you understand what "first-class entity" means, grasshopper. Who modded this insightful?
The GP was talking about the way arrays are handled as addresses with pointers instead of being more ingrained in the language.
That has nothing to do with classes at all. You saw the word "class" and you went into overdrive, did you not?
Both of your comments show how strong the language is. And why I would never *ever* like to read a few thousand lines of your code. YUCK.
Take off new line *especially*? WTF?
l = [int(x.strip()) for x in l.strip().split(',')]
Are you really trying to say I should prefer this to the methods available in Java? I (*and* you) would easily understand a piece of code doing the same thing in Java.
I've had the same issues with C/C++ and a lot of other languages though. Most Java programs run fine on the latest JDK, and some run up to 20/30% faster on JDK 1.6. There are a few frameworks that try to cope with this problem, such as the OSGi framework for Java and assemblies for .NET. But it is a complex problem and I don't think *any* language has solved this problem completely. It might not even be completely solvable, come to think of it.
My code compiled against 1.2 still runs fine in JDK 1.6 though. So I cannot call that horrible or anything.
It's almost open source, so the problem with the Sun classes may disappear, as you can create a duplicate of the functionality somewhere else.
Obviously you still have the problem that the code may be 1) not well tested 2) unreadable 3) tied in with other code.
Avoiding code in the sun.* packages should definitely be the default. Make sure you have an IDE that can filter them out of your auto-complete lists.
Try the checkstyle plugin in Eclipse (for Java of course). And thank god there is auto-complete, clean and fix :)