I admit that I don't fully understand all the issues of swap, but here are some observations:
Generational garbage collectors, such as the one used in the JVM, screw up swap. It seems like there is a conflict between what the OS is trying to do with swap and what the JVM is trying to do with GC. I would rather let the GC win in this so the application runs fast.
Memory is cheap these days. I just started up OpenOffice, widely regarded as a hog of a process, and it takes 27mb. Huge, you say? That's about $2 worth of memory these days.
Does swap actually make things faster? The concensus these days seems to be "if your process is swapping you need more memory." Therefore... why swap?
It seems like the place where swap makes sense is that you can have a bunch of non-running/idle processes all "running" at the same time, while swapped out. I'm not sure how much this is needed on modern machines with > 1gb of ram.
People on/. love to talk about all things to do with MS vs. Linux and yet this poor suffering discussion only generates a couple dozen comments. I will add in something about the best real-world way to use Wine: Codeweaver's Crossover Office. It really works. Microsoft Office 2000 works perfectly in Crossover Office, and Office XP is completely usable for day-to-day work. It is amazing how well these things work and how well they integrate into the Linux desktop. Wine is the foundation of it, and it works. If it works this well now, how will it be by the end of the year? They also have a cascade effect, in that if they solve a bug in MS Office, that might also solve bugs in many other untested applications. I have noticed that the unsupported apps work better and better. Wine is a relevant and cool project. The thing that might make it irrelevant is that Linux desktop software options are catching up. I think that OpenOffice.org is already better than MS Office in many respects.
Effective Java
on
Hardcore Java
·
· Score: 5, Interesting
by Joshua Bloch is the best programming book I have ever read in any language. It goes into detail about exceptions and how to use them properly. It sounds like it covers everything which this book covers, but it's brilliant. It's also a thin book. It is the only computer science book I have ever read multiple times, and gotten something out of it every time I have read it. Anyone else who has read this book is invited to chime in here...
Rolling your own desktop distro is a subtle thing. Sun has always been a hardware company, and so they have always had control over the hardware their OS runs on. They have never yet had to deal with the tangled mess that is PC hardware, with conflicts, obscure devices, and all the rest of it. And the one thing missing most from the Sun Java desktop is Java. Where are all the Java apps? Sun should be all over this: "The Sun Java Desktop is a collection of apps, protocols and file formats which let you run your desktop environment anywhere that runs Java. We have partnered with Suse and Redhat to provide an environment which we certify is Sun Java Desktop compatible, but any Java 1.4 environment will work." What about doing that? But Sun is not doing that. Is anyone? Yes. But if you look at their website, you see that they are backed by American investors, but not Sun! What's going on? Sun should buy them and make that the cornerstone of the Java desktop.
In future, if Sun really wants something it can call the Sun Java Desktop, it would have all the applications in Java, and a Java runtime which is perfectly integrated into the OS, like OS X's Java environment.
The problem with looking for solutions to the energy problem is that there isn't a solution. There are a whole bunch of small solutions that, when added up together, will be the solution. One plant producing 500 barrels/day is 1/10,000th of the solution. One thousand such plants is 1/10th of the solution. Add in a few nuclear reactors, some solar panels, wind turbines, more efficient cars, biodiesel, 100% electric cars with lithium batteries, telecommuting, maybe even a Segway, and it starts adding up to a solution to the energy problem. If we did all of those things in parallel, within a few years, OPEC would be sweating and we would not have to spend billions of dollars a year on oil, and then billions more on trying to keep our oil suppliers stable and friendly.
I also hear people say "the oil industry has too much power here for anything to change." This is also the wrong view. Sure, the oil industry does have a lot of power, but the result of their machinations is that our entire economy is dependent on a commodity which we must import from politically unstable and hostile parts of the world which are far away. There are plenty of other powerful industries in the US that have nothing to do with oil that must see this as a hazardous situation, one which should be remedied by moving the US to having multiple energy options to choose from, including cost-competitive domestic solutions. Is the oil industry in the US more powerful than all the other non-oil industries? I don't think so.
"This being 2004, you should know not to open a file from an untrusted source." WRONG! This is exactly the mindset that has resulted in the security problems that plague computers today. Operating environments should have the ability to fully contain and isolate any process. Operating environments should have the ability to run hostile code with complete safety. The smart thing to do is to start regarding ALL code as hostile. One side effect of that is that failures of non-hostile code will be contained, too, making for a more reliable system.
How can such a goal be attained? There are many ways available now. The most obvious one is a VM system with security policies, such as the JVM. That's not the only one, though. Another method is a capabilities-based system, so when a process starts, it has only a defined set of capabilities to work with. OpenBSD has a similar, but more limited system called systrace. The TrustedBSD project and SELinux have similar aims, and SELinux is being integrated into mainstream Linux distros. Another way to run untrusted things is with user-mode Linux, which I believe is integrated with Linux 2.6
The editor is right, though, that on currently-used systems like OSX and MS Windows, you have to be careful what you click on. But the problem is that we have come to accept that as "the way things are", when there is no reason for that to be the case. You should be able to run hostile code, see what it does, laugh at it, and delete it without any harm. The technology to do that exists, and has existed for years, but we have come to accept broken products and systems that don't allow that.
Exchange is probably MS' best and most important product that has no Open Source equivalent. I am aware of Kolab and some other works-in-progress, but none of these are even close to Exchange yet. Exchange has more than its fair share of security problems, but what it does, it does well. Now with Connector being released GPL, that will have two consequences: The free downloadable version of Evolution will be able to use Exchange's features, and hopefully other OS tools like Koffice/Kmail will pick up those abilities, too. Also, having an open source client side might help them in getting an open source server side move faster. Now I just wish that Evolution would be properly integrated with KDE. They are doing it with OOo...
I'm a full-time desktop Linux user, and not just for coding, but for every aspect of business, so all this stuff matters to me. This week is going to be a great week in Desktop Linux: Suse 9.1 and Crossover 3 are both coming out at about the same time, and both are huge improvements over what came before.
I'm not yet sure of the details, but do you remember the scene in Brasil where the Central Services guys were "fixing" the HVAC systems in the appartment, and then Tuttle comes in on his zip line and does a cross-over between the sewer output line and the suit air input line and then the two Central Services guys suits fill up with sewage and explode? We need to somehow cross-connect the fax spammers with 419 scammers so they work on eachother.
I've used their word processor. It works very well. It loads fast and feels very polished. The only problem is that it still uses a proprietary file format (MS.doc format) which is exactly what I'm trying to avoid. There are so many great tools out there for manipulating XML programatically; it seems like XML is the only viable option right now for creating files which will be searchable and transformable for years to come. If I use some strange proprietary format like MS.doc, basically my files are locked up and someone else has the keys to them. Not a good position to be in.
However, the next version of TextMaker is supposed to support.sxw (OpenOffice.org) formats, as is KOffice 1.4. Then we will have a truly portable document format for use with cross-platform office software.
What good is a directional beam if it hits some radio-reflective object and bounces somewhere else? Also, even if the beam is 99% directional, sensitive or very close receivers could still pick up the 1% that leaks. A security system that is 99% effective is not much better than a system which is 10% effective. Without solid encryption and authentication built-in to the protocol, directional broadcasting is useless. With solid encryption and authentication built-in to the protocol, directional broadcasting doesn't add anything.
The one place where this could have some good security uses is for undetectable transmission, which is probably interesting to the military.
Of course, directional broadcasting has a whole set of real benefits, such as getting more bandwidth by allowing more transmitters in the same region, minimizing interference, minimizing radiation output, etc. But to call this a security feature? I guess the "everything good is a security feature" is the parallel to "everything bad is terrorist" idea which seems popular lately.
I use OOo all the time, and find it to be very powerful, more powerful than MS Office in many ways. If it had better KDE integration it would be perfect... fortunately, KDE integration is on its way, and it will be fantastic, giving serious competition to anything else (including OSX) on the desktop. It makes sense to drop Koffice from Knoppix, and it will make even more sense once the KDE/OOo integration is ready.
Yesterday I was helping my friend get set up with a Java application she needs. She's a fairly experienced Windows user. She needed to install the Sun JRE 1.4 to get this thing to work. She needed help from me, and a bit of exploring on java.sun.com, to find out which is the right file for her to download to get this thing working. If Sun wants MS Windows users to upgrade to be able to run Java apps easily and painlessly, there needs to be a big link right on java.sun.com saying "Windows users, click here to download Java for your system." Such a link does not exist and the user has to figure out "do I want NetBeans (no), do I need a SDK (no), do I need J2ME (no), do I need the JRE (yes)". Ridiculous. How is she supposed to understand the Java technology family in order to know how to navigate this? Is Sun trying to kill Java? Is there some secret Microsoft/SCO/Knights Templar conspiracy infiltrating in Sun? Ok, probably no on that last one...
And now the Sun "Java" Desktop, which presumably comes with Java built in, but does that mean you can double-click on a.jar file and your app starts up? No. You have to write a shell script, or add an icon with a command like "java -classpath foo.jar..." to get it to work.
Sun, what are you thinking?
Maybe expecting Sun to move from the world of big servers, where expecting users to write a shell script is perfectly acceptable, to the world of desktops, where users should be able to do everything just by clicking in an obvious place and without having to understand the difference between an ELF file and a JAR, is too much to ask.
An authentic Java desktop would be, in my opinion, one in which all the work gets done in Java. That means a Java office suite, a Java window manager, a Java file explorer. This is completely doable, and Java is a fantastic environment for doing those things (I know, you will flame me saying Java sucks, Java is slow, etc, sorry, that isn't true anymore). I would love to see such a desktop environment, and it would have fantastic security and portability advantages. A real Java-based OS is the only thing that has a real chance of competing with Linux, I believe (ok, I will get majorly flamed for that, but it's true).
I think that if Sun is serious about this, the way forward is:
Partner up with Suse/Novell. They are going to win in the desktop by producing a real-world desktop distro. In fact I'm using it right now and it's great.
Open-source Java, or give enough assistance to existing OS java projects such as Kaffe and GNU Classpath to make them viable real-world usable Java implementations. Sun thinks this is irrelevant; it's not. Getting something truly open source is a key step to making something ubiquitous.
Develop some real-world Java apps. As someone else on/. said, J2ee is becoming the Cobol of our age. That is sad because Java could be so much more than a server system. How about developing a word processor and spreadsheet in Java? By doing this, they would give people a reason to use Java, and they would also really find out what the limitations and flaws of Swing are, and maybe they would fix them.
Ok, that's enough ranting, sure to stir up many heated flames about how much Java sucks, and Scott hasn't called me anytime recently to ask for business advice, so I'll leave off here.
Skype for Linux is on its way, and it can't be here soon enough for me. If you look on their site in the jobs section, they are looking for a Linux/QT programmer. Older versions of Skype ran under Wine, but apparently they are now doing some code relocation/decryption thing that breaks with Wine, so we just have to wait for the official Linux client. And if it's Qt, it will look good and integrate nicely with Suse 9.1 which is also almost here...
If you would like to read a book that is a brilliant mix of great writing, science, philosophy, conspiracies and the Transit of Venus, as witnessed in Africa, I highly recommend "Mason & Dixon" by Thomas Pynchon. It does take a while to get through it and you need to have Google and a dictionary handy to understand some of the more obscure references in it, but it is both funny and sad and very worth reading. Basically, Mason and Dixon, the two cartographers behind the Mason Dixon Line, are dispatched to various places in the world to make various observations, and the most interesting is their assignment to South Africa to observe the Transit.
Of course there is malware for Windows and Linux because both are written in unsafe languages which manipulate memory directly and often cast objects to (void *) and use containers (pointers to regions of memory) which don't know their own size. I know I've posted on this before here on/., but as long as we use those tools, we're going to have those problems. It's interesting that he doesn't have any examples of Java malware, for instance.
This is a hydrocarbon with a bunch of floruine in it. It's similar enough to many general anesthetics that I would imagine it could be used as one in its gas form. If the molecule is small enough to be absorbed in the lungs and carried through the blood-brain barrier, I'm guessing it would be a GA. Not something I want in my computer case.
I have pre-ordered my Suse 9.1 and I can't wait to try it out. From what I can see in these screen shots, it looks like it is far ahead of MS Windows in many areas of usability. It looks like Steve Jobs should be taking a close look at some of the features. It also looks like it comes with more built-in stuff (things like OCR) than either MS Windows or OSX. Hopefully Amazon will hurry up with my package.
Gregory Nemitz, president of Orbital Development, said: "The MoonCrash Project would probably be attractive to some bored rich guy, who is tired of playing with his radio-controlled model airplanes and wants to move up to the next level."
Wow, that doesn't seem like the right tone for someone trying to sell a $6mil unnecessary luxury product.
If software patents had existed thirty or forty years ago, there would be patents on the for loop, linked lists, the subroutine, use of buffers, storing text in the form of a sequence of bytes (an invention called a "string"), use of APIs, re-use of code in a form called a "library", the process of merging multiple files of executable code together known as "linking", and many other things, and the process of taking a file containing "source code" and turning it into executable (a novel invention known as "compiling"). All of these things would be patented. Non-commercial software could not exist because there's just no way to write a program without strings, lists and loops, unless your program just consists of a long list of NOOP instructions. But of course NOOP could have been patented too. In fact all these inventions are much more patentable than the stuff that is getting patented now, because when these software constructs were developed, they really were completely novel and non-obvious, unlike things like one-click ordering.
It has already been mentioned here, but tcpflow is great. It captures a tcp session into a file. Sometimes if I just can't figure out what is going on with a web application, the best way to really see what the client is getting in real life is to tcpflow it and capture a session. Sometimes that shows up the bug. Tcpdump is good but it puts out individual packets which are hard to piece together. Tcpflow does all of that. Of course, it isn't so useful for protocols other than tcp.
"Security is about patches." That statement implies the belief that security flaws are inevitable, an inherent part of having software. This simply isn't true. We should not accept such thinking. If a product doesn't have security holes the day it is released, it still won't have security holes a thousand years from now, patches or not. The question is, how do we ship products without holes? The reasons we have security holes in products are not because developers are stupid or careless, or because the business side of the company wants to ship the product now. No, the reason we have holes is because we're still using horrible software development tools which make security problems almost inevitable. Humans just can't think like C compilers and if we write a long enough program in plain old C, we end up with buffer overflows and lack of bounds checking on things. If we used safer tools like Java, which don't have buffers and which store data in structures which know their own size (collections), the vast majority of vulnerabilities would never even be created. If a user sends malicious input to a Java process, we know that no matter how broken the Java is, that malicious input can't stomp on memory and be executed, no matter what, because the JVM and the bytecode verifier don't allow it to. That is the kind of assurance that software should have.
It is always possible to make security problems at the design level, like forgetting to check an account balance before allowing a withdrawal in bank software, but humans are very good at thinking in those ways, and those kinds of problems are rare.
This being/., the topic of Ogg must come up. I would like to buy an Ipod but there's just no way I will buy any player that can't play Ogg files. I guess if they get Linux running on it well enough and they have a fast enough Ogg decoder, I should be able to play Ogg files on the Ipod? Then it would be worth it.
Slashdot Mirror
- Generational garbage collectors, such as the one used in the JVM, screw up swap. It seems like there is a conflict between what the OS is trying to do with swap and what the JVM is trying to do with GC. I would rather let the GC win in this so the application runs fast.
- Memory is cheap these days. I just started up OpenOffice, widely regarded as a hog of a process, and it takes 27mb. Huge, you say? That's about $2 worth of memory these days.
- Does swap actually make things faster? The concensus these days seems to be "if your process is swapping you need more memory." Therefore... why swap?
It seems like the place where swap makes sense is that you can have a bunch of non-running/idle processes all "running" at the same time, while swapped out. I'm not sure how much this is needed on modern machines with > 1gb of ram.----------
WAP software
-----------
WAP software
-----------
WAP software
In future, if Sun really wants something it can call the Sun Java Desktop, it would have all the applications in Java, and a Java runtime which is perfectly integrated into the OS, like OS X's Java environment.
-----------
WAP Apache software
I also hear people say "the oil industry has too much power here for anything to change." This is also the wrong view. Sure, the oil industry does have a lot of power, but the result of their machinations is that our entire economy is dependent on a commodity which we must import from politically unstable and hostile parts of the world which are far away. There are plenty of other powerful industries in the US that have nothing to do with oil that must see this as a hazardous situation, one which should be remedied by moving the US to having multiple energy options to choose from, including cost-competitive domestic solutions. Is the oil industry in the US more powerful than all the other non-oil industries? I don't think so.
------------
Create a WAP server
How can such a goal be attained? There are many ways available now. The most obvious one is a VM system with security policies, such as the JVM. That's not the only one, though. Another method is a capabilities-based system, so when a process starts, it has only a defined set of capabilities to work with. OpenBSD has a similar, but more limited system called systrace. The TrustedBSD project and SELinux have similar aims, and SELinux is being integrated into mainstream Linux distros. Another way to run untrusted things is with user-mode Linux, which I believe is integrated with Linux 2.6
The editor is right, though, that on currently-used systems like OSX and MS Windows, you have to be careful what you click on. But the problem is that we have come to accept that as "the way things are", when there is no reason for that to be the case. You should be able to run hostile code, see what it does, laugh at it, and delete it without any harm. The technology to do that exists, and has existed for years, but we have come to accept broken products and systems that don't allow that.
---------
WAP news
I'm a full-time desktop Linux user, and not just for coding, but for every aspect of business, so all this stuff matters to me. This week is going to be a great week in Desktop Linux: Suse 9.1 and Crossover 3 are both coming out at about the same time, and both are huge improvements over what came before.
-------
WAP news
I'm not yet sure of the details, but do you remember the scene in Brasil where the Central Services guys were "fixing" the HVAC systems in the appartment, and then Tuttle comes in on his zip line and does a cross-over between the sewer output line and the suit air input line and then the two Central Services guys suits fill up with sewage and explode? We need to somehow cross-connect the fax spammers with 419 scammers so they work on eachother.
-----------
WAP news
Maybe I should create a wireless interface to it...
However, the next version of TextMaker is supposed to support .sxw (OpenOffice.org) formats, as is KOffice 1.4. Then we will have a truly portable document format for use with cross-platform office software.
-------------
Create a WAP server
The one place where this could have some good security uses is for undetectable transmission, which is probably interesting to the military.
Of course, directional broadcasting has a whole set of real benefits, such as getting more bandwidth by allowing more transmitters in the same region, minimizing interference, minimizing radiation output, etc. But to call this a security feature? I guess the "everything good is a security feature" is the parallel to "everything bad is terrorist" idea which seems popular lately.
-----------
Create a WAP server
----------
Create a WAP site
And now the Sun "Java" Desktop, which presumably comes with Java built in, but does that mean you can double-click on a .jar file and your app starts up? No. You have to write a shell script, or add an icon with a command like "java -classpath foo.jar ..." to get it to work.
Sun, what are you thinking?
Maybe expecting Sun to move from the world of big servers, where expecting users to write a shell script is perfectly acceptable, to the world of desktops, where users should be able to do everything just by clicking in an obvious place and without having to understand the difference between an ELF file and a JAR, is too much to ask.
An authentic Java desktop would be, in my opinion, one in which all the work gets done in Java. That means a Java office suite, a Java window manager, a Java file explorer. This is completely doable, and Java is a fantastic environment for doing those things (I know, you will flame me saying Java sucks, Java is slow, etc, sorry, that isn't true anymore). I would love to see such a desktop environment, and it would have fantastic security and portability advantages. A real Java-based OS is the only thing that has a real chance of competing with Linux, I believe (ok, I will get majorly flamed for that, but it's true).
I think that if Sun is serious about this, the way forward is:
Ok, that's enough ranting, sure to stir up many heated flames about how much Java sucks, and Scott hasn't called me anytime recently to ask for business advice, so I'll leave off here.
-------Create a WAP server
----------
Create a WAP server
------------
Create a WAP server
----------
Create a WAP server
------------
Create a WAP server
----------
Create a WAP server
Wow, that doesn't seem like the right tone for someone trying to sell a $6mil unnecessary luxury product.
--------
WAP server software
--------
Create a WAP server
--------
WAP hosting
-------
Create a WAP server
It is always possible to make security problems at the design level, like forgetting to check an account balance before allowing a withdrawal in bank software, but humans are very good at thinking in those ways, and those kinds of problems are rare.
---------
Create a WAP server
--------
WAP hosting