If open source software could compete, why has Apache's market share been steadily losing ground since Windows Server 2003 was released? If you follow the current trends, in a little under two years Microsoft will have the lead.
Extrapolating long term predictions out of short term trends does not usually yield accurate results. For instance, you could say exactly the same thing about the period of June to December in 2001. Indeed, IIS market share has yet to even reach regain its peak in March 2002, and if you look at the current running figures for November, you'll notice that Microsoft's market share has actually dropped a little.
Making deals with large domain-parking companies like Go Daddy couldn't have hurt Microsoft's market share much, either.
This is opinion, not a fact. Some of the highest traffic and bandwidth sites on the Internet are using standard IIS and SQL installations. And they not only work, but work well at what they do. If what you believe was true, these sites could not possibly do what they do using MS's products.
I didn't say IIS couldn't run websites; I merely said that IIS was mediocre. It provides basic functionality, integrates well with.NET and ASP, and is configured through a GUI. Problems only occur when you start to push the envelope, to do something more advanced than just serving up pages, such as URL rewriting, or reverse proxies to legacy applications. You also run into problems when you want to use non-Microsoft languages. Whilst Apache is flexible enough to adapt to any scenario, IIS is limited to a subset of problems.
IIS clearly is usable, but only if you're planning to construct a relatively straightforward site using Microsoft technologies. Unfortunately for Microsoft, the majority of websites don't fall under that umbrella. If IIS wants to capture more market share, it really needs to become more adaptable.
MS is also vastly easier to deploy than most solutions. MS has done a good job of making easy to use administration tools and easing the lives of server administrators. This is also an area that a lot of competive products, especially in the Open Source world, need to pay more attention to. Not everyone has the *nix skills and even people like myself that do, would rather make a few clicks than edit several configuration files.
I'm not so sure about that. These days it's rather easy to install a Linux distribution on a server, as many distributions have a number of preset install settings to choose from. Put in the boot CD, select "install web server", then after it's finished use a web browser to access webadmin, so that you can customize Apache without touching a command line.
Of course, some people are more comfortable with the things they know, so I suspect some people just plump for the known option; i.e. Windows. But the majority of the server market, as IBM notes, is tied up by Linux and the other BSDs, with Apache still the dominant webserver by a large margin. My point is simply that Microsoft is unlikely to gain market share, unless they expand out of their niche.
Because until that statement, the discussion was about Windows vs Linux. You seemed to be arguing against Windows on the basis that IIS-on-Windows is crappy compared to Apache-on-Linux. I was trying to point out that Apache-on-Windows was a better comparison.
I admit was deviating back to the original topic a little, which concerns Linux dominating the server side market. It's my opinion that if Microsoft want to increase their market share, they'd have to make a better argument than, "It runs all your favourite server software, just like Linux, and it's more expensive."
So on Linux you can install MySql or Oracle, and install Apache, right? Guess what. On Windows you can install MySql or Oracle, and install Apache. Its not hard.
Uh, yes, you can install alternatives to Microsoft products. I'm not sure how that invalidates my assertion that Microsoft's products tend to be somewhat mediocre offerings.
Besides, much as you (and I for that matter) may not like it, SQL server runs some pretty massive installations very well.
Funny you should mention that. SQL Server's scaling capabilities are currently the major bottleneck that my company is having trouble with.
And IIS works better than Apache does for serving up small amounts of random data - I can wander over to our hosted Windows server, pull up the IIS console, and set up a bandwidth limited virtual server on a specific domain in about 3 minutes.
Sure, and I didn't say that IIS was useless, just that it was mediocre, especially considering the price Microsoft puts on it. It's not designed to be particularly feature-packed or flexible, and it seems that if Microsoft want to make gains in the server space, they need something a little better.
For 95% of the tasks out there, almost any software works these days.
You just can't compete with that marketing budget. Not when people with no technical knowledge make the purchasing decisions. Not only is Microsoft encouraging you to buy their own products, but the thousands of other tech companies that bring in billions of dollars of revenue each year by selling products that make Microsoft's broken bloated trash usable are encouraging you to buy Microsoft so that you'll need their software to fix them. In 5 years, Microsoft will have the same stranglehold on the server market that they have on the desktop today. Ironically, they may blow the desktop market with Vista.
If open source server software, such as Linux, cannot compete with Microsoft's marketing budget, why does Apache continue to hold its substantial majority in the webserver market? Whilst server operating system statistics are somewhat hard to find, webserver market share is well documented. If your theory is correct, then one would expect IIS to be making similar gains against Apache, yet Microsoft have failed to put any significant dent in Apache's market share in the decade or so they've been around.
Either a) web servers are managed by more tech-savvy staff than most other servers, b) Microsoft has been uninterested in expanding its webserver market share, or c) Microsoft's marketing doesn't matter as much as you think it does.
Neither a) nor b) seem very likely to me. Web servers are not significantly more difficult to configure than your average database system or mail server, so it's unlikely that the companies managing these systems are considerably more tech-savvy than those managing other server-based systems. Likewise, it seems somewhat unlikely that Microsoft is uninterested in the web-server market, considering its size and relative visibility. Thus, the only conclusion I can draw is that you've overestimated the effectiveness of Microsoft's marketing department.
I find it hard to believe Microsoft could overturn the substantial Linux server market share in 5 years, when it hasn't been able to effectively challenge Apache in 10.
So I will ask, give us even one example of something that Linux is capable of that Windows is not capable of doing.
I think the grandparent is incorrect in asserting that Linux can do things that Windows cannot handle. With enough time and effort, you can get Windows to do the same things Linux can, or at least a mediocre approximation thereof. I use Linux at home, and Windows XP at my workplace, and often things that I could do easily in Linux, are hard and unwieldy in Windows. For example, getting a decent command line involves installing Cygwin; virtual desktops involve a third-party hack which doesn't work too well; installing software requires me to follow installation instructions instead of apt-getting packages - the list goes on.
And don't get me started on the limitations of IIS and SQL Server compared to their open source brethren. You really have to admire Microsoft for making so much money from mediocre products.
Your explanation of an onion-router is interesting. It sounds like you're talking about a distributed key decryption, where each node on the route does part of the decryption, but only the ends can see the plain text. I'm not sure how that's any more effective at protecting data, only at stopping surveillance from making a connection between the end points.
Essentially yes. Onion routing rather effectively masks the connection between the origin of a connection and its end source. As far as I know, this sort of scheme is only inherently vulnerable to two attacks. First is if the majority of onion routers are compromised, which is rather obvious and somewhat unlikely flaw. The second is in timing attacks, which essentially look for correlation between the times users request pages, and the times servers receive requests. However, the latter attack is difficult to pull off unless you're monitoring a particular server for access by a particular person, and thus unsuitable for widespread monitoring.
Onion routing is a connection hiding technology, and does little to protect data. It does prevent your ISP from seeing your data (as its encrypted before it passes into the routing network), which may count for something, but it doesn't protect responses from the end server from being intercepted. However, data security is a relatively well understood field, and is a generally simpler problem to solve than effective anonymity over a packet switching network.
But an awful lot of sites that peoply surf using anonymizers aren't using encrypted streams. I'd expect them to be very traceable by checksum-correlation.
True. But as with security, the majority of the populous is always going to be under-protected. However, projects such as Torpark are making it easier to achieve effective anonymity with relatively little technological experience, so perhaps the number of people using effective tools to mask their identity will increase over the next few years.
It doesn't take much of a hashing function to boil down a response packet to a 32-64 bit integer. Watch the packets going into the anonymizer, calculating their hash functions on the body (not the IP header.)
Watch for the outcoming packets, and where they're destined. 99% of the time it'll give a point-to-point address map.
I'm sorry, but I fail to understand how this would work. How would hashing the packets help? The packets sent to and from the client would be encrypted; the packets sent to and from the server would be unencrypted (or encrypted through a different system, such as SSL). How would hashing a packet help you match unencrypted packets to encrypted packets when the encryption key is unknown?
I don't think you have a very good idea of how modern onion routers work. When I send a packet through a network like Tor, it is encrypted by the client in such a way that only a specific order of onion routers can decrypt it. In order to match a response packet to a client, you would need control over all of those routers, as no individual one has the information necessary to match the original response packet to its destination.
In theory you could route through multiple anonymizers working as a peer network, but as long as you can keep an eye on the end points, the "in between" of the anonymizer network is just noise.
There's no "in theory" about it. Any decent anonymising system does that already, and uses a rather ingenious encryption scheme to prevent the attacks to describe.
Anonymizer routing can still be tracked, it just takes more work and some high-powered address correlation hardware. Or a simple but massive gate array, looking for data checksum correlations between streams entering and leaving an anonymizer.
You make it sound a relatively simple affair, when in reality it's anything but. In order to track down a request passed through a standard onion-routed network, you really need to have have access to both ends of the data stream. If a government wished to reliably track who was visiting website X, they'd need access to several pieces of information. The first thing they'd need would be the complete access logs from all the ISPs and Internet cafes operating in the country, or perhaps a suitably random sample if deterrence is their only goal. The second thing they'd need is the server logs of website X (or at least the logs of the hosting company), which is further complicated if website X is being hosted in another country. They'd then need a way to sift through all this information in order to correlate the access logs with the server logs, and even if they managed that herculean task, there would still be a large margin of error involved.
In other words, trying to work out who visited a website, assuming that a large number of visitors are using reasonably good anonymising networks, is impractical even for a large IT-literate government.
Things become somewhat better when a single user Y is targeted, and you have an idea what sort of websites they would be visiting. This would narrow the solution space down considerably, and you'd have a good chance at discovering whether or not user Y was visiting website X. Again, assuming that you can get access to the server logs website X, and assuming that the server logs are kept at all, and are genuine.
A totalitarian government might be able to get around these restrictions through a variety of means, such as mandatory keyloggers, or by instilling enough fear into its citizens to discourage dissent. But a democratic government that has a largely fair judicial system doesn't stand much of a chance.
It is harder to use. I know, not once you've LEARNED it, but look: if I go into a store, buy a digital camera, and ship it to a person who has never used a computer before, they can read the installation instructions and get it working. You can't do the same with Linux, because the non-techie documentation level just isn't there.
I plugged my Cannon A60 into my computer, and Ubuntu popped up a dialog telling me it had noticed my camera, and if would I like to download my pictures. I didn't have to put in a CD and run through a multi-step installation process, as I did with Windows. Under Linux, it all just worked first time, so if anything, Linux is far easier to use in this regard than Windows is.
That said, Windows does have a greater range of hardware support than Linux does, but the user interface for the digital cameras that are supported is far better and more intuitive than Windows, so I think you've picked a bad example.
On Windows or Mac I can shove in a CD, it can automount, and install the software itself; Linux software generally does not do that.
That's true; with most Linux software, you don't have to mess around with CDs or installation wizards. On Ubuntu for instance, you select "Add/Remove..." from the "Applications" menu, choose the application you want, then press the install button, and that's it. The downloading and installing of the application is handled automatically.
So again, you put forward an example where Linux is easier to use than Windows. And again, the problem is not with ease of use, but with support. Some of the applications a user might want to use (such as the latest computer game, or an application like Photoshop), are simply not available for Linux. But that's a different problem entirely.
Even a universal installer would help. Click an icon in the menu labelled "Install new software", and it says "Do you... ( ) have a.deb file? ( ) have an.rpm file? ( ) have a.tar.gz file? ( ) have an installation CD? ( ) want to download it from the internet?" Do that, and suddenly my wife can download a file (say a.deb) from a web site, run the installer, and get it going -- even if I've never told her what to do.
It's true that consistency is a big problem for Linux (though in Ubuntu, all one has to do to install a deb is to double-click on it, then hit 'install' on the resulting dialog box). Progress is being made in that direction, but it'll a year or more before we start to reap most of the benefits, I'd imagine.
True, but on the other hand the problems seem all to lie with apt-get upgrades from Dapper. AFAIK, Microsoft have no equivalent functionality that allows them to upgrade their OS whilst its still running.
That said, I've had my fair share of Edgy problems. One of the new native wireless drivers conflicted with my ndiswrapper driver I used on Dapper, preventing either driver from working properly. It took me the best part of a day to find the cause of the problem. Apart from that, however, everything upgraded without a hitch, and Beryl with AIGLX and the NVidea beta drivers is surprisingly stable.
The key thing is ad-blocking functionality that automatically downloads effective rules. I don't have the time to maintain a list of regular expressions, so Filterset.G is a significant timesaver for me. I'd be interested to know if anything like that has been developed for Opera; last I checked, it hadn't. Another set of functionality I use in Firefox is the various web developer extensions, such as Firebug and EditCSS. Also PasswordMaker, which again isn't available for Opera, and TorButton, when I'm feeling paranoid.
Those are all the things I can think of currently, though I'm sure there's a few more. Whilst Opera no doubt suits some people down to the ground, Firefox has the edge in features and (because I use GNOME) desktop integrations. Another problem with Opera is that the Linux version had some smooth scrolling issues back with version 8.0, though they may have sorted that out in 9.0.
That's not what I've found. The memory consumption issues of Firefox 1.5.x have still not been dealt with. The Firefox process I'm using right now has been running since yesterday afternoon. Using the Task Manager, I can see that Firefox is taking up 593 MB of RAM. I've heard that this can be caused by bad extensions, so I didn't install any. Furthermore, I heard that Firefox's caching sometimes uses a lot of memory, so I completely disabled it.
I'm always intrigued by these comments. There's barely a time at work when a Firefox window isn't open in the background, I have numerous extensions installed, and having over two dozen tabs open is not particularly unusual for me; however, Firefox has never even come close to using up that much RAM on any machine I've worked on, even when I have that amount of memory to spare. Even the huge pages the new Slashdot comment system produces doesn't raise my RAM usage very far over the 100M mark, and the majority of that is likely caching.
I wonder why Firefox seems to use up so much memory for some people, whilst others get away with relatively little. Did you have any plugins installed that might have been the cause of this problem?
5. You ignore the negative effects that the drug barons have on society (organised crime of other kinds).
Wait, surely this is an argument for legalising drugs? Criminals can profit from drug trafficking because its illicit nature allows them to have extremely high margins with none of the governmental oversight usually associated with the pharmaceutical business. If one could buy heroin or cocaine from the local chemist, organised crime gangs would be quickly priced out of the market by large pharma corporations. Doubtless there'd still be some money to be made from tax-dodging, but this would be a fraction of the market.
So the question is whether you believe that the disadvantages of legalising drug use outweighs the advantages of significantly reducing the profits of organised crime.
The idea that 'people should be allowed to do what they want with their own body' is wrong. It's wrong because it's based on the premise that we don't owe anything to society. No matter how independant you might think you are, you still owe a huge debt to society, and its ancestry.
By that argument, suicide should be made illegal, since you're depriving society of your future contributions. Besides, paying back debts to society is exactly what taxes are for. If drug use increases our debt, then we should pay increased taxes; the high tax on cigarettes and alcohol is an obvious precedent.
Arguing that we shouldn't be able to do what we want with our own bodies, implies that our bodies are not entirely our property. I'm not sure I particularly like the idea of this.
Stay you speaker of the Dread Cthulhu, tell me more: By the Mad Arab's death, I know I am Thane of Innsmouth, But how, of Arkham? the Thane of Arkham liues A prosperous Gentleman: And to be King, Stands not within the prospect of belief, No more then to be Arkham. Say from whence You held word with the Great Priest, or why Upon this blasted Heath you stop our way And tell of things meant not for mortal ears Speak, I charge you.
To be fair, Firefox has had extensions supporting this functionality for almost as long as Opera, as I recall using Session Saver back when Firefox was Phoenix. This is just the first time this functionality has been built into the browser by default.
Let me throw it back in your face - since when is it your right to use someone else's creation in violation of the terms to which they agreed to sell it to you?
Generally speaking, this occurs when the terms violate local law, be it a competition law, a privacy law, or a reverse engineering law, a fair use law, or whatever other law might apply to the distribution of digital information.
Indeed, by choosing to sell your product in a sovereign country, are you not implicitly agreeing to adhere to their laws? It seems a little rich to claim that someone is violating the terms under which you sell, when you yourself are violating the terms that allow you to sell in the first place.
Then on the other side you get the grunt programming. Writing the same app thats been written a hundred times before, nothing new or particularly challenging besides dealing with the customer/managment/deadlines.
Repetition is a job for a computer, not a programmer. There should be no need for so-called 'grunt programming', as that is a symptom derived from bad programming. I recognise that there are many programmers of little talent in the world; but there are relatively few programmers able to practise their craft with sufficient competance. There's no white collar and blue collar programmers; just skilled and unskilled.
As coding goes from being an elite craft that few can do into a blue color job that few bother to want to do
If you're suggesting that programming is no longer as difficult as it was, then this would rather imply, a) programming viable applications is easier now than it was, or b) people are far more intelligent than they were. Neither explanation seems very likely. As far as I can see, programming is still a craft that very few people are capable of doing with any competance.
However, if you are suggesting that it has lost some of its 'elite' image, then perhaps you are right.
If you're doing something a fascist regime might find objectionable, you can just run off a live CD, so there will be no evidence of your browsing activity on your computer at all.
But as you point out, if you're targetted by a fascist regime (and the US, for the most part, is not), then your best hope is not to be targetted in the first place, which is where anonymising services actually do some good.
Perhaps I'm underestimating the number of shallow consumers in the music industry. However, this would still imply that the recording industry would still decline in size somewhat, with those who buy music because of how it sounds gradually turning to means of distribution that would more accurately cater to their needs. In such a world, the RIAA would be left to cater for those who only buy music to further an image, regardless of how the music actually sounds, whilst any artist that appeals to geniune musical taste will find more lucrative deals elsewhere.
Or perhaps I'm again falling into the classic economic trap of assuming that a market will make perfectly rational decisions.
Slashdot Mirror
Extrapolating long term predictions out of short term trends does not usually yield accurate results. For instance, you could say exactly the same thing about the period of June to December in 2001. Indeed, IIS market share has yet to even reach regain its peak in March 2002, and if you look at the current running figures for November, you'll notice that Microsoft's market share has actually dropped a little.
Making deals with large domain-parking companies like Go Daddy couldn't have hurt Microsoft's market share much, either.
I didn't say IIS couldn't run websites; I merely said that IIS was mediocre. It provides basic functionality, integrates well with .NET and ASP, and is configured through a GUI. Problems only occur when you start to push the envelope, to do something more advanced than just serving up pages, such as URL rewriting, or reverse proxies to legacy applications. You also run into problems when you want to use non-Microsoft languages. Whilst Apache is flexible enough to adapt to any scenario, IIS is limited to a subset of problems.
IIS clearly is usable, but only if you're planning to construct a relatively straightforward site using Microsoft technologies. Unfortunately for Microsoft, the majority of websites don't fall under that umbrella. If IIS wants to capture more market share, it really needs to become more adaptable.
I'm not so sure about that. These days it's rather easy to install a Linux distribution on a server, as many distributions have a number of preset install settings to choose from. Put in the boot CD, select "install web server", then after it's finished use a web browser to access webadmin, so that you can customize Apache without touching a command line.
Of course, some people are more comfortable with the things they know, so I suspect some people just plump for the known option; i.e. Windows. But the majority of the server market, as IBM notes, is tied up by Linux and the other BSDs, with Apache still the dominant webserver by a large margin. My point is simply that Microsoft is unlikely to gain market share, unless they expand out of their niche.
I admit was deviating back to the original topic a little, which concerns Linux dominating the server side market. It's my opinion that if Microsoft want to increase their market share, they'd have to make a better argument than, "It runs all your favourite server software, just like Linux, and it's more expensive."
Uh, yes, you can install alternatives to Microsoft products. I'm not sure how that invalidates my assertion that Microsoft's products tend to be somewhat mediocre offerings.
Funny you should mention that. SQL Server's scaling capabilities are currently the major bottleneck that my company is having trouble with.
Sure, and I didn't say that IIS was useless, just that it was mediocre, especially considering the price Microsoft puts on it. It's not designed to be particularly feature-packed or flexible, and it seems that if Microsoft want to make gains in the server space, they need something a little better.
Unfortunately I seem to handle the remaining 5% :(
If open source server software, such as Linux, cannot compete with Microsoft's marketing budget, why does Apache continue to hold its substantial majority in the webserver market? Whilst server operating system statistics are somewhat hard to find, webserver market share is well documented. If your theory is correct, then one would expect IIS to be making similar gains against Apache, yet Microsoft have failed to put any significant dent in Apache's market share in the decade or so they've been around.
Either a) web servers are managed by more tech-savvy staff than most other servers, b) Microsoft has been uninterested in expanding its webserver market share, or c) Microsoft's marketing doesn't matter as much as you think it does.
Neither a) nor b) seem very likely to me. Web servers are not significantly more difficult to configure than your average database system or mail server, so it's unlikely that the companies managing these systems are considerably more tech-savvy than those managing other server-based systems. Likewise, it seems somewhat unlikely that Microsoft is uninterested in the web-server market, considering its size and relative visibility. Thus, the only conclusion I can draw is that you've overestimated the effectiveness of Microsoft's marketing department.
I find it hard to believe Microsoft could overturn the substantial Linux server market share in 5 years, when it hasn't been able to effectively challenge Apache in 10.
I think the grandparent is incorrect in asserting that Linux can do things that Windows cannot handle. With enough time and effort, you can get Windows to do the same things Linux can, or at least a mediocre approximation thereof. I use Linux at home, and Windows XP at my workplace, and often things that I could do easily in Linux, are hard and unwieldy in Windows. For example, getting a decent command line involves installing Cygwin; virtual desktops involve a third-party hack which doesn't work too well; installing software requires me to follow installation instructions instead of apt-getting packages - the list goes on.
And don't get me started on the limitations of IIS and SQL Server compared to their open source brethren. You really have to admire Microsoft for making so much money from mediocre products.
Essentially yes. Onion routing rather effectively masks the connection between the origin of a connection and its end source. As far as I know, this sort of scheme is only inherently vulnerable to two attacks. First is if the majority of onion routers are compromised, which is rather obvious and somewhat unlikely flaw. The second is in timing attacks, which essentially look for correlation between the times users request pages, and the times servers receive requests. However, the latter attack is difficult to pull off unless you're monitoring a particular server for access by a particular person, and thus unsuitable for widespread monitoring.
Onion routing is a connection hiding technology, and does little to protect data. It does prevent your ISP from seeing your data (as its encrypted before it passes into the routing network), which may count for something, but it doesn't protect responses from the end server from being intercepted. However, data security is a relatively well understood field, and is a generally simpler problem to solve than effective anonymity over a packet switching network.
True. But as with security, the majority of the populous is always going to be under-protected. However, projects such as Torpark are making it easier to achieve effective anonymity with relatively little technological experience, so perhaps the number of people using effective tools to mask their identity will increase over the next few years.
I'm sorry, but I fail to understand how this would work. How would hashing the packets help? The packets sent to and from the client would be encrypted; the packets sent to and from the server would be unencrypted (or encrypted through a different system, such as SSL). How would hashing a packet help you match unencrypted packets to encrypted packets when the encryption key is unknown?
I don't think you have a very good idea of how modern onion routers work. When I send a packet through a network like Tor, it is encrypted by the client in such a way that only a specific order of onion routers can decrypt it. In order to match a response packet to a client, you would need control over all of those routers, as no individual one has the information necessary to match the original response packet to its destination.
There's no "in theory" about it. Any decent anonymising system does that already, and uses a rather ingenious encryption scheme to prevent the attacks to describe.
You make it sound a relatively simple affair, when in reality it's anything but. In order to track down a request passed through a standard onion-routed network, you really need to have have access to both ends of the data stream. If a government wished to reliably track who was visiting website X, they'd need access to several pieces of information. The first thing they'd need would be the complete access logs from all the ISPs and Internet cafes operating in the country, or perhaps a suitably random sample if deterrence is their only goal. The second thing they'd need is the server logs of website X (or at least the logs of the hosting company), which is further complicated if website X is being hosted in another country. They'd then need a way to sift through all this information in order to correlate the access logs with the server logs, and even if they managed that herculean task, there would still be a large margin of error involved.
In other words, trying to work out who visited a website, assuming that a large number of visitors are using reasonably good anonymising networks, is impractical even for a large IT-literate government.
Things become somewhat better when a single user Y is targeted, and you have an idea what sort of websites they would be visiting. This would narrow the solution space down considerably, and you'd have a good chance at discovering whether or not user Y was visiting website X. Again, assuming that you can get access to the server logs website X, and assuming that the server logs are kept at all, and are genuine.
A totalitarian government might be able to get around these restrictions through a variety of means, such as mandatory keyloggers, or by instilling enough fear into its citizens to discourage dissent. But a democratic government that has a largely fair judicial system doesn't stand much of a chance.
I plugged my Cannon A60 into my computer, and Ubuntu popped up a dialog telling me it had noticed my camera, and if would I like to download my pictures. I didn't have to put in a CD and run through a multi-step installation process, as I did with Windows. Under Linux, it all just worked first time, so if anything, Linux is far easier to use in this regard than Windows is.
That said, Windows does have a greater range of hardware support than Linux does, but the user interface for the digital cameras that are supported is far better and more intuitive than Windows, so I think you've picked a bad example.
That's true; with most Linux software, you don't have to mess around with CDs or installation wizards. On Ubuntu for instance, you select "Add/Remove..." from the "Applications" menu, choose the application you want, then press the install button, and that's it. The downloading and installing of the application is handled automatically.
So again, you put forward an example where Linux is easier to use than Windows. And again, the problem is not with ease of use, but with support. Some of the applications a user might want to use (such as the latest computer game, or an application like Photoshop), are simply not available for Linux. But that's a different problem entirely.
It's true that consistency is a big problem for Linux (though in Ubuntu, all one has to do to install a deb is to double-click on it, then hit 'install' on the resulting dialog box). Progress is being made in that direction, but it'll a year or more before we start to reap most of the benefits, I'd imagine.
True, but on the other hand the problems seem all to lie with apt-get upgrades from Dapper. AFAIK, Microsoft have no equivalent functionality that allows them to upgrade their OS whilst its still running.
That said, I've had my fair share of Edgy problems. One of the new native wireless drivers conflicted with my ndiswrapper driver I used on Dapper, preventing either driver from working properly. It took me the best part of a day to find the cause of the problem. Apart from that, however, everything upgraded without a hitch, and Beryl with AIGLX and the NVidea beta drivers is surprisingly stable.
The key thing is ad-blocking functionality that automatically downloads effective rules. I don't have the time to maintain a list of regular expressions, so Filterset.G is a significant timesaver for me. I'd be interested to know if anything like that has been developed for Opera; last I checked, it hadn't. Another set of functionality I use in Firefox is the various web developer extensions, such as Firebug and EditCSS. Also PasswordMaker, which again isn't available for Opera, and TorButton, when I'm feeling paranoid.
Those are all the things I can think of currently, though I'm sure there's a few more. Whilst Opera no doubt suits some people down to the ground, Firefox has the edge in features and (because I use GNOME) desktop integrations. Another problem with Opera is that the Linux version had some smooth scrolling issues back with version 8.0, though they may have sorted that out in 9.0.
I'm always intrigued by these comments. There's barely a time at work when a Firefox window isn't open in the background, I have numerous extensions installed, and having over two dozen tabs open is not particularly unusual for me; however, Firefox has never even come close to using up that much RAM on any machine I've worked on, even when I have that amount of memory to spare. Even the huge pages the new Slashdot comment system produces doesn't raise my RAM usage very far over the 100M mark, and the majority of that is likely caching.
I wonder why Firefox seems to use up so much memory for some people, whilst others get away with relatively little. Did you have any plugins installed that might have been the cause of this problem?
It doesn't have the broad range of functionality Firefox gets from its extensions. That's a fairly big downside for me.
Wait, surely this is an argument for legalising drugs? Criminals can profit from drug trafficking because its illicit nature allows them to have extremely high margins with none of the governmental oversight usually associated with the pharmaceutical business. If one could buy heroin or cocaine from the local chemist, organised crime gangs would be quickly priced out of the market by large pharma corporations. Doubtless there'd still be some money to be made from tax-dodging, but this would be a fraction of the market.
So the question is whether you believe that the disadvantages of legalising drug use outweighs the advantages of significantly reducing the profits of organised crime.
By that argument, suicide should be made illegal, since you're depriving society of your future contributions. Besides, paying back debts to society is exactly what taxes are for. If drug use increases our debt, then we should pay increased taxes; the high tax on cigarettes and alcohol is an obvious precedent.
Arguing that we shouldn't be able to do what we want with our own bodies, implies that our bodies are not entirely our property. I'm not sure I particularly like the idea of this.
Plastikfox is a theme for Firefox that mimics the default Plastik theme. Openoffice also has a Qt interface that blends into KDE fairly flawlessly.
To be fair, Firefox has had extensions supporting this functionality for almost as long as Opera, as I recall using Session Saver back when Firefox was Phoenix. This is just the first time this functionality has been built into the browser by default.
Oho! That I have to try! Do you know if it allows for transparent PNGs referenced in the CSS background-image property?
I was dreading the inevitable process of trying to get a new CSS design working in IE 6; but hopefully now I don't have to :)
Generally speaking, this occurs when the terms violate local law, be it a competition law, a privacy law, or a reverse engineering law, a fair use law, or whatever other law might apply to the distribution of digital information.
Indeed, by choosing to sell your product in a sovereign country, are you not implicitly agreeing to adhere to their laws? It seems a little rich to claim that someone is violating the terms under which you sell, when you yourself are violating the terms that allow you to sell in the first place.
Repetition is a job for a computer, not a programmer. There should be no need for so-called 'grunt programming', as that is a symptom derived from bad programming. I recognise that there are many programmers of little talent in the world; but there are relatively few programmers able to practise their craft with sufficient competance. There's no white collar and blue collar programmers; just skilled and unskilled.
If you're suggesting that programming is no longer as difficult as it was, then this would rather imply, a) programming viable applications is easier now than it was, or b) people are far more intelligent than they were. Neither explanation seems very likely. As far as I can see, programming is still a craft that very few people are capable of doing with any competance.
However, if you are suggesting that it has lost some of its 'elite' image, then perhaps you are right.
If you're doing something a fascist regime might find objectionable, you can just run off a live CD, so there will be no evidence of your browsing activity on your computer at all.
But as you point out, if you're targetted by a fascist regime (and the US, for the most part, is not), then your best hope is not to be targetted in the first place, which is where anonymising services actually do some good.
Perhaps I'm underestimating the number of shallow consumers in the music industry. However, this would still imply that the recording industry would still decline in size somewhat, with those who buy music because of how it sounds gradually turning to means of distribution that would more accurately cater to their needs. In such a world, the RIAA would be left to cater for those who only buy music to further an image, regardless of how the music actually sounds, whilst any artist that appeals to geniune musical taste will find more lucrative deals elsewhere.
Or perhaps I'm again falling into the classic economic trap of assuming that a market will make perfectly rational decisions.