I think your logic is correct. If our hypothetical researcher had to provide his own facility, do his own system administration or hire someone else to do it, pay his own electrical bills including the cost of cooling, manage his own network, etc.... then cloud computing is economically far better. But since this person will be working in a context where he or she is only responsible for the hardware and software costs of the physical machines, buying machines is much cheaper.
People at all ranks of the company are often "a fall guy for things you did not cause", but they don't get a multi-million dollar severance package when they get fired.
The gold in "golden parachute" is piss on the people who didn't make the right friends at a country club so they could get a cushy job and walk away rich no matter whether they got the job right or totally fucked it up.
Motorola was filing lawsuits against other Android manufacturers. By making this acquisition, Google will probably prevent that from happening. That's one win.
At the beginning of this year Google had less than a thousand patents (which is especially odd since they got their initial venture capital on the strength of founder Larry Page's patent on their search technology). Then they bought 1000 patents from IBM. Motorola Mobility has 5000 US patents and 1500 more pending, plus over 10000 more and more pending in other countries. Google's patent portfolio is not at Apple and Microsoft's level yet, but they just went from a welter weight in the patent wars to the middle weight division.
I never used one of the commercial products for virtualization. We were constrained to $0 for the software budget for virtualization, so we toyed with VirtualBox and KVM on Linux. Our development machine was not on a UPS, and over the course of a few weeks we had the occasional 20 second power outage in the building. Some of our VirtualBox images were corrupted by the outage, the KVM images were not, and that was enough to put us onto KVM for production (even though our production servers are of course on UPS with a backup generator). It's possible whatever problem we had with VirtualBox has been fixed in more recent versions or that we misconfigured the storage settings, I don't know. But KVM was more reliable without any tweaking right out of the box, so we went with that.
KVM supports live migration and live storage migration, although we have not used either feature. The virt-manager GUI you can use with KVM is easy enough - create, clone, start, stop, change settings, and view and interact with the virtual machines all with clicks in the GUI. I'm sure VMWare has earned its impressive reputation, but free is always nice. Good luck.
Borders and Barnes & Noble employ cashiers, people to stock shelves, and janitors and have thousands of retail outlets which require running water, air conditioning, heating, and local taxes in addition to the cost of running a website and distribution warehouses. Amazon only has the website and distribution warehouses, their business costs per book sold were dramatically lower. Borders probably did have special pricing deals with the publishing companies, but they needed to net 40% or more profit per book to cover their operating expenses. Amazon only needs 10% per book to cover their operating expenses.
When XP failed to configure a device correctly the machine was left in a usable state. I just downloaded the correct driver, install, reboot, and use the device. If the device with a problem was a network driver, I just downloaded the driver on another machine, burned it to CD, put CD in the machine I'm setting up, install from there, reboot, and use the device.
When Debian failed to detect my hard drives and aborted the installation, or Suse failed to configure my wireless card, or Fedora had no sound, or Ubuntu crashed at boot with no usable display, I would have a completely unusable system. I would need to use a separate computer to research the issue. Often I could not find a solution, period. When I did find a solution, it often involved booting a liveCD, changing a bunch of settings by hand so Linux could boot to a usable state, booting back into the broken installation, and then running more commands to attempt to fix it.
I am not saying that Windows is better. It is not. I'm just saying that my experience is totally different from mcgrew. With PCLinuxOS, Aptosid Linux, and Ubuntu 10.10 I finally had the super easy install that put Linux head and shoulders past Windows for ease of installation. Before that Windows was unquestionably easier.
I'm running 11.04 and gave up after a month of trying to have Unity run for more than 10 minutes at a time without losing all keyboard and mouse input until I restart the X server. I reverted to their Gnome-Classic interface. I listened to an interview with the Compiz developer that Canonical (the company that funds Ubuntu) hired to help with Unity. The guy was intelligent, polite, and humble and clearly trying very hard to make a good product. But Canonical clearly pushed this out before it was ready. I really like most of Unity's features, but I can't exactly recommend it to people when it doesn't fucking work.
Your experience differs from mine. I have had hardware detection issues with most Linux installations every year from 2001 to 2007. It was only in 2008 or so that I started to consistently have my video card, wireless card, and sound work consistently right out of the gate with most distributions and even now I can't get Debian Squeeze or OpenSUSE to install. By comparison Windows is tedious to install but I haven't had an installation fail or fail to properly configure hardware since Windows 95.
I'm happy for you and the wonderful experience you've had. I haven't been so lucky.
The problem with optimizing for platform on the PC is that the platform varies. Some of your optimizations for the ATI Radeon HD 4000 series won't carry over to the 5000 series. Some of your optimizations for Windows XP won't carry over to Windows 7. Some of your optimizations for 64 bit don't carry over to the big chunk of the population still running 32 bit. And in five years someone is going to buy a new PC with Intel integrated graphics that have roughly equivalent processing power to the nVidia GPU you spent $300 on in 2009, but all of your nVidia GT-specific optimizations won't help with that Intel chip.
The console is a fixed target - optimize for that, and you are guaranteed to give decent performance to more than ten million potential customers.
That's why your PC has three times the processing power, ten times the memory, and ten times the GPU power compared to your PS3 or Xbox360 but only gets 50% better graphics.
The article says the Wii 2 will have four USB ports. But if you have 25 GB of game data, the throughput from USB 2 might be too slow for you to run a game off it. That's just my wild guess, though. (Granted that 98% of Slashdot posts can be suffixed with, "That's just my wild guess, though.")
Our profits are small, which is one of the reasons we avoid SQL Server. But further back in the discussion, someone wrote "You're not going anywhere running windows unless you had shit loads of cash behind you." and several other comments above and below mine in the discussion have similar statements. Even 2% of gross revenue spent buying Microsoft's most expensive software product is not "shitloads of cash".
I work for a small company that does about $1 million in gross revenues per year. We run Windows on all of the workstations and laptops, Windows Server on all of the servers, and everyone has Office Professional. We spend around $3000 on Microsoft software per year - 0.3% of gross revenue.
If we ran SQL Server instead of PostgreSQL and had MSDN subscriptions for Microsoft developer environments we would probably be spending maybe $20,000 per year. That's 2% of gross revenue, harsher but still a relatively small expense.
You might like Aptosid, it's based on Debian unstable. You have to use command prompts for your software updates, so that can be annoying, but other than that it's excellent.
Most but not all of the problems are from installation. Apple and the major PC vendors that sell computers with Windows usually do an excellent job making sure your installation is complete and all of your devices are compatible and configured correctly before they sell you the computer. If you bought a pre-configured Linux computer from a trustworthy source, you would probably have a similar experience.
Of course, that doesn't help with all of the Windows applications you would not be able to run at all, or would have to do lots of extra work to run under Wine. It also doesn't help if you want to perform a major upgrade on your operating system. To be fair, most people with Microsoft operating systems don't do a major operating system upgrade, they just wait until the PC is old and then replace it entirely. But if you bought a pre-built PC that runs (for the sake of argument) Fedora 13 beautifully and then a few crucial features break when you upgrade to Fedora 14, your frustration is understandable.
I don't recommend Linux to people unless they're really interested in tackling learn-as-you-go projects or I had a really easy time getting it to work for myself. So far the latter happens very rarely. I play plenty of games with an ATI video card, the ATI proprietary device driver for it, and Wine - but it's not easy enough to set up and maintain that I would recommend it. ( For example, every time Ubuntu releases a kernel update I have to uninstall my proprietary driver, reboot, install the new kernel, reboot, and reinstall the driver. If I forget, my display is hosed and I have to ssh into the machine remotely to uninstall the driver. I could just use the open source driver for my card, but then Starcraft 2 goes from 25 fps to a slide show. )
The ability to switch back to the Ubuntu Classic interface is getting removed in the next release. So that buys me maybe a year, but then what do I do when I want the latest versions of Firefox, Libre Office, Wine, and so forth along with an interface I prefer? I either install it from source myself or switch distributions.
But further, what do I tell people that want to try Linux? "Use Ubuntu, it's great. Just avoid the default interface." What kind of endorsement is that?
Most importantly, these complaints about Unity are very straightforward and easy to understand:
1. I used to switch virtual desktops with one mouse click. In Unity, it takes three.
2. I used to launch my programs by three mouse clicks. Menu -> Sub-Menu -> Item I want. In Unity, if I want a Firefox browser, media player, or word processor I launch with one click. For anything else, like games, development environment, alternative browser, alternative media player, etc... I have to click the upper left button, and type in a search based on the name.
3. I used to switch between running instances of programs with one mouse click. In Unity, switching between items on the menu panel is easy, for everything else it's completely unintuitive.
With Microsoft Windows Computers, iPhones, iPads, Android Phones, and just about any other consumer tech product, it gets heavy advertisement in print, television, and internet sites, and you can walk into a store and see the product and get a demonstration or overview from a sales rep.
Linux distributions don't have that, there has never been enough sales of Linux PCs from big retailers for them to consistently stock them on the store shelves and give them a prominent display position. None of the major vendors, including Canonical and Red Hat, have the money to make big television, internet, or print advertisements. Linux has primarily spread through curiosity and word of mouth. When a significant portion of your user base has their word of mouth recommendation change from Ubuntu to something else, there is no way in hell your rate of adoption is going to speed up. Unless the remainder of your users positively fall in love with your new design and start shouting from the rooftops how awesome they think it is, your rate of adoption is going to slow down.
I'm not opposed to a radical change to user interface. To me Ubuntu Classic feels pretty generic, and I don't mind something newer. But in Unity it takes me three clicks of a mouse to switch virtual desktops instead of one. I can't fit all the programs I routinely use on their menu bar, so instead of navigating quickly through menus I have to constantly type in the search box. For someone that does web surfing, watches movies, and runs a word processor maybe Unity is an improvement. For me it slows me down significantly, and I'm not going to recommend a Linux distribution to users if I have to qualify my advice with, "It works really well just as long as you avoid the default user interface!" That's absurd.
If you use auto-login, the login screen doesn't let you change your desktop environment. You have to log in, search Unity for Login Screen, click Unlock, enter your password, change your setting from "Ubuntu" to "Ubuntu Classic", close that, then logout and log back in.
I'm really, really irritated at Unity. Ubuntu was my default recommendation for Linux newbies that asked me what they should try, and it has been for two years. I find the Unity user interface very unintuitive. I'm not going to tell people to try Ubuntu if complete removal of the Ubuntu Classic is on the roadmap.
We're not talking about saving $15, we're talking about saving hundreds of dollars or getting a nicely supported $1000 Linux laptop that has the same performance as a $500 Windows 7 laptop. A lot of people aren't able or willing to spend the extra money, and your attitude towards them isn't going to help spread Linux.
I want Linux to be more common. For every thousand new users you're going to get a handful that help other newbies and a few that contribute to more Linux-related open source projects. The more of us there are, the better it will be for all of us. Being rude is not helpful.
Slashdot Mirror
I think your logic is correct. If our hypothetical researcher had to provide his own facility, do his own system administration or hire someone else to do it, pay his own electrical bills including the cost of cooling, manage his own network, etc.... then cloud computing is economically far better. But since this person will be working in a context where he or she is only responsible for the hardware and software costs of the physical machines, buying machines is much cheaper.
People at all ranks of the company are often "a fall guy for things you did not cause", but they don't get a multi-million dollar severance package when they get fired.
The gold in "golden parachute" is piss on the people who didn't make the right friends at a country club so they could get a cushy job and walk away rich no matter whether they got the job right or totally fucked it up.
Motorola was filing lawsuits against other Android manufacturers. By making this acquisition, Google will probably prevent that from happening. That's one win.
And while Apple and Microsoft have patent lawsuits against Motorola, Motorola has in turn filed strong counter-suits. http://www.patentlyo.com/patent/2011/08/googles-purchase-of-motorola-mobility.html
At the beginning of this year Google had less than a thousand patents (which is especially odd since they got their initial venture capital on the strength of founder Larry Page's patent on their search technology). Then they bought 1000 patents from IBM. Motorola Mobility has 5000 US patents and 1500 more pending, plus over 10000 more and more pending in other countries. Google's patent portfolio is not at Apple and Microsoft's level yet, but they just went from a welter weight in the patent wars to the middle weight division.
Yes, picking "KVM" for "Kernel Virtual Machine" when most people think of it as "Keyboard, Video, Mouse" was absurd.
I never used one of the commercial products for virtualization. We were constrained to $0 for the software budget for virtualization, so we toyed with VirtualBox and KVM on Linux. Our development machine was not on a UPS, and over the course of a few weeks we had the occasional 20 second power outage in the building. Some of our VirtualBox images were corrupted by the outage, the KVM images were not, and that was enough to put us onto KVM for production (even though our production servers are of course on UPS with a backup generator). It's possible whatever problem we had with VirtualBox has been fixed in more recent versions or that we misconfigured the storage settings, I don't know. But KVM was more reliable without any tweaking right out of the box, so we went with that.
KVM supports live migration and live storage migration, although we have not used either feature. The virt-manager GUI you can use with KVM is easy enough - create, clone, start, stop, change settings, and view and interact with the virtual machines all with clicks in the GUI. I'm sure VMWare has earned its impressive reputation, but free is always nice. Good luck.
Unity is not the same thing as GNOME 3. Ubuntu 11.04 runs Unity on top of parts of GNOME, but not full GNOME 3.
I like Ubuntu's Unity interface, but it was so buggy and crash-happy on my hardware that I had to ditch it anyway.
Borders and Barnes & Noble employ cashiers, people to stock shelves, and janitors and have thousands of retail outlets which require running water, air conditioning, heating, and local taxes in addition to the cost of running a website and distribution warehouses. Amazon only has the website and distribution warehouses, their business costs per book sold were dramatically lower. Borders probably did have special pricing deals with the publishing companies, but they needed to net 40% or more profit per book to cover their operating expenses. Amazon only needs 10% per book to cover their operating expenses.
When XP failed to configure a device correctly the machine was left in a usable state. I just downloaded the correct driver, install, reboot, and use the device. If the device with a problem was a network driver, I just downloaded the driver on another machine, burned it to CD, put CD in the machine I'm setting up, install from there, reboot, and use the device.
When Debian failed to detect my hard drives and aborted the installation, or Suse failed to configure my wireless card, or Fedora had no sound, or Ubuntu crashed at boot with no usable display, I would have a completely unusable system. I would need to use a separate computer to research the issue. Often I could not find a solution, period. When I did find a solution, it often involved booting a liveCD, changing a bunch of settings by hand so Linux could boot to a usable state, booting back into the broken installation, and then running more commands to attempt to fix it.
I am not saying that Windows is better. It is not. I'm just saying that my experience is totally different from mcgrew. With PCLinuxOS, Aptosid Linux, and Ubuntu 10.10 I finally had the super easy install that put Linux head and shoulders past Windows for ease of installation. Before that Windows was unquestionably easier.
I like Unity. Just as soon as someone can make it work for more than ten minutes on my machine, I'll use it.
I'm running 11.04 and gave up after a month of trying to have Unity run for more than 10 minutes at a time without losing all keyboard and mouse input until I restart the X server. I reverted to their Gnome-Classic interface. I listened to an interview with the Compiz developer that Canonical (the company that funds Ubuntu) hired to help with Unity. The guy was intelligent, polite, and humble and clearly trying very hard to make a good product. But Canonical clearly pushed this out before it was ready. I really like most of Unity's features, but I can't exactly recommend it to people when it doesn't fucking work.
Your experience differs from mine. I have had hardware detection issues with most Linux installations every year from 2001 to 2007. It was only in 2008 or so that I started to consistently have my video card, wireless card, and sound work consistently right out of the gate with most distributions and even now I can't get Debian Squeeze or OpenSUSE to install. By comparison Windows is tedious to install but I haven't had an installation fail or fail to properly configure hardware since Windows 95.
I'm happy for you and the wonderful experience you've had. I haven't been so lucky.
The problem with optimizing for platform on the PC is that the platform varies. Some of your optimizations for the ATI Radeon HD 4000 series won't carry over to the 5000 series. Some of your optimizations for Windows XP won't carry over to Windows 7. Some of your optimizations for 64 bit don't carry over to the big chunk of the population still running 32 bit. And in five years someone is going to buy a new PC with Intel integrated graphics that have roughly equivalent processing power to the nVidia GPU you spent $300 on in 2009, but all of your nVidia GT-specific optimizations won't help with that Intel chip.
The console is a fixed target - optimize for that, and you are guaranteed to give decent performance to more than ten million potential customers.
That's why your PC has three times the processing power, ten times the memory, and ten times the GPU power compared to your PS3 or Xbox360 but only gets 50% better graphics.
The article says the Wii 2 will have four USB ports. But if you have 25 GB of game data, the throughput from USB 2 might be too slow for you to run a game off it. That's just my wild guess, though. (Granted that 98% of Slashdot posts can be suffixed with, "That's just my wild guess, though.")
Our profits are small, which is one of the reasons we avoid SQL Server. But further back in the discussion, someone wrote "You're not going anywhere running windows unless you had shit loads of cash behind you." and several other comments above and below mine in the discussion have similar statements. Even 2% of gross revenue spent buying Microsoft's most expensive software product is not "shitloads of cash".
I work for a small company that does about $1 million in gross revenues per year. We run Windows on all of the workstations and laptops, Windows Server on all of the servers, and everyone has Office Professional. We spend around $3000 on Microsoft software per year - 0.3% of gross revenue.
If we ran SQL Server instead of PostgreSQL and had MSDN subscriptions for Microsoft developer environments we would probably be spending maybe $20,000 per year. That's 2% of gross revenue, harsher but still a relatively small expense.
You might like Aptosid, it's based on Debian unstable. You have to use command prompts for your software updates, so that can be annoying, but other than that it's excellent.
If you don't like it, take Android 2.x (which is open) and run with it.
Most but not all of the problems are from installation. Apple and the major PC vendors that sell computers with Windows usually do an excellent job making sure your installation is complete and all of your devices are compatible and configured correctly before they sell you the computer. If you bought a pre-configured Linux computer from a trustworthy source, you would probably have a similar experience.
Of course, that doesn't help with all of the Windows applications you would not be able to run at all, or would have to do lots of extra work to run under Wine. It also doesn't help if you want to perform a major upgrade on your operating system. To be fair, most people with Microsoft operating systems don't do a major operating system upgrade, they just wait until the PC is old and then replace it entirely. But if you bought a pre-built PC that runs (for the sake of argument) Fedora 13 beautifully and then a few crucial features break when you upgrade to Fedora 14, your frustration is understandable.
I don't recommend Linux to people unless they're really interested in tackling learn-as-you-go projects or I had a really easy time getting it to work for myself. So far the latter happens very rarely. I play plenty of games with an ATI video card, the ATI proprietary device driver for it, and Wine - but it's not easy enough to set up and maintain that I would recommend it. ( For example, every time Ubuntu releases a kernel update I have to uninstall my proprietary driver, reboot, install the new kernel, reboot, and reinstall the driver. If I forget, my display is hosed and I have to ssh into the machine remotely to uninstall the driver. I could just use the open source driver for my card, but then Starcraft 2 goes from 25 fps to a slide show. )
The ability to switch back to the Ubuntu Classic interface is getting removed in the next release. So that buys me maybe a year, but then what do I do when I want the latest versions of Firefox, Libre Office, Wine, and so forth along with an interface I prefer? I either install it from source myself or switch distributions.
But further, what do I tell people that want to try Linux? "Use Ubuntu, it's great. Just avoid the default interface." What kind of endorsement is that?
Most importantly, these complaints about Unity are very straightforward and easy to understand:
1. I used to switch virtual desktops with one mouse click. In Unity, it takes three.
2. I used to launch my programs by three mouse clicks. Menu -> Sub-Menu -> Item I want. In Unity, if I want a Firefox browser, media player, or word processor I launch with one click. For anything else, like games, development environment, alternative browser, alternative media player, etc... I have to click the upper left button, and type in a search based on the name.
3. I used to switch between running instances of programs with one mouse click. In Unity, switching between items on the menu panel is easy, for everything else it's completely unintuitive.
You summed it up perfectly.
With Microsoft Windows Computers, iPhones, iPads, Android Phones, and just about any other consumer tech product, it gets heavy advertisement in print, television, and internet sites, and you can walk into a store and see the product and get a demonstration or overview from a sales rep.
Linux distributions don't have that, there has never been enough sales of Linux PCs from big retailers for them to consistently stock them on the store shelves and give them a prominent display position. None of the major vendors, including Canonical and Red Hat, have the money to make big television, internet, or print advertisements. Linux has primarily spread through curiosity and word of mouth. When a significant portion of your user base has their word of mouth recommendation change from Ubuntu to something else, there is no way in hell your rate of adoption is going to speed up. Unless the remainder of your users positively fall in love with your new design and start shouting from the rooftops how awesome they think it is, your rate of adoption is going to slow down.
I'm not opposed to a radical change to user interface. To me Ubuntu Classic feels pretty generic, and I don't mind something newer. But in Unity it takes me three clicks of a mouse to switch virtual desktops instead of one. I can't fit all the programs I routinely use on their menu bar, so instead of navigating quickly through menus I have to constantly type in the search box. For someone that does web surfing, watches movies, and runs a word processor maybe Unity is an improvement. For me it slows me down significantly, and I'm not going to recommend a Linux distribution to users if I have to qualify my advice with, "It works really well just as long as you avoid the default user interface!" That's absurd.
If you use auto-login, the login screen doesn't let you change your desktop environment. You have to log in, search Unity for Login Screen, click Unlock, enter your password, change your setting from "Ubuntu" to "Ubuntu Classic", close that, then logout and log back in.
I'm really, really irritated at Unity. Ubuntu was my default recommendation for Linux newbies that asked me what they should try, and it has been for two years. I find the Unity user interface very unintuitive. I'm not going to tell people to try Ubuntu if complete removal of the Ubuntu Classic is on the roadmap.
We're not talking about saving $15, we're talking about saving hundreds of dollars or getting a nicely supported $1000 Linux laptop that has the same performance as a $500 Windows 7 laptop. A lot of people aren't able or willing to spend the extra money, and your attitude towards them isn't going to help spread Linux.
I want Linux to be more common. For every thousand new users you're going to get a handful that help other newbies and a few that contribute to more Linux-related open source projects. The more of us there are, the better it will be for all of us. Being rude is not helpful.
I upgraded from 10.10 to 11.04, and the option is not on my login screen. I have to log in, go to System -> Login Screen, and change it there.