Who doesn't shut steam down after they are done playing?
People who have the vast majority of their games from the last x years in Steam and don't mind the few dozen MB of RAM it takes up in the background for the time savings of just being able to launch any game in two clicks?
People who use the Steam IM system to see when their friends are online and playing a game they might want to join?
And who doesn't close all programs before shutting down?
Who manually closes any of their autorun-at-startup programs before shutting down? Or really anything? Make sure anything open you care about is saved, Start, Shutdown.
What the hell yourself? Are you one of those crazy people who thinks empty RAM is a good thing? Before I had a SSD the Steam client took about 15 seconds to reach the login screen, then another 20-30 before my game list was available. Now these are in the 5 and 10-15 second ranges respectively, but still more than just right-clicking the Steam icon in my tray or pinned to my taskbar and going right to the game I want. Unless you're really tight on RAM or want to prevent people from playing your Steam-purchased games on your computer without logging in first, just treating it as a constantly running service is a good common choice.
There was a long-standing bug in Steam which made offline mode unreliable for anyone who shuts their computer down regularly.
When you'd shut down Windows, Steam would apparently just sit around and fail to heed the warnings the OS gave it until it would be forced to close. This made it not properly save the offline mode token, so if the next time you launched the client was without an internet connection, you were SOL.
If you closed Steam before shutting down it worked fine, but since this wasn't common knowledge and you don't usually plan internet outages at the consumer level as far as an average user cares it didn't work.
The actual cause of the bug was identified publicly by a user a few months ago and shortly followed by a Steam client update which resolved the problem. At this point offline mode works exactly as expected.
Re:Seems like a perfect application for BitTorrent
on
Black Mesa Released
·
· Score: 2
They do have a torrent, just they hosted the tracker on GameUpdates which pretty much crumbled under the weight. Why they didn't use one of the many public open high-capacity trackers I do not know.
My company had a need for a file server. I was the most technically knowledgeable in the company, so the task fell on me. It was being done on a zero budget using a spare computer, so I chose Linux. The rest you can guess from there.
You mean the OS that, by default, blocks you from running content that isn't blessed by Apple? Yes, you can download apps from sources that aren't the App Store - but they still have to be signed, otherwise, it either will refuse to run or lie to you and say that the app is "damaged" and you should "drag it to the trash."
Try not to make shit up, it makes your argument look like the rant of a crazy fanboy to anyone who knows better regardless of whether other parts are valid.
First, it's not anything about content blessed by Apple. Any registered Apple developer gets a key with which they can sign all their apps as they desire. Apple approval only matters for their App Store. Gatekeeper can restrict a computer to App Store apps only, but that is not the default setting and must be chosen by the user.
Admittedly the developer registration costs $99/year to maintain. This should not be a problem for any commercial developer, even at the smallest scale, but obviously may be a barrier for freeware developers. Anyone developing for iOS (officially) is already a member as well, making it effectively free for them. I would note as well that a key for Microsoft's comparable Authenticode system costs three times as much per year, though it is not currently enforced by the operating system in many user-visible ways (the most notable one being the dialog before launching downloaded EXEs).
Second, where the hell are you getting this idea that it would lie and say the application is damaged? That happens if the application IS signed but does not match what's expected based on the signature. An unsigned app in the default configuration gets a dialog that says it's been blocked by the computer's current security policy without in any way implying that such apps are necessarily bad. The dialog practically comes out and tells the user where to change this as well, the few who couldn't figure it out from that point are exactly the kind of user this is intended to protect.
When the Gatekeeper feature is switched off (I haven't upgraded to 10.8 yet so I can't confirm or deny your claim of a warning of dire consequences while doing this), the warning on unsigned applications changes to one which is practically identical to the one Windows displays on downloaded apps. It only shows once, then goes away forever for that application.
Third, sending diagnostic and usage data is disabled by default. It asks to send such data after certain kinds of failures, and if you select "Don't Ask Me Again" before clicking yes then sure it will enable itself, but that was your choice as the user. The data collected (and sent if enabled) is available to the user from the "Console" app, same as any other logs on the system. Remember, these are still computers on which the user can run any apps they please, inspect the hard drive as they please, and change any certificates they please. What's sent by the computer can be inspected by the user with ease. If there was a difference between what Apple claims, what Console shows, and what's actually being logged and transmitted it'd be trivial to see and call them out on it.
The only time Apple knows what non-AppStore apps you have on your Mac is if there's a crash which you choose to report to Apple (or have previously chosen to send all to them) and the app is actually running at the time of the crash. Even in these cases it's not guaranteed, as the OS can decide that other apps weren't relevant to the problem in which case it doesn't bother logging them. I just looked at the five crash logs I have since installing the OS, and not a single one has even a process list in it. All that is logged is the application that crashed, how it crashed, information about its memory usage, what libraries it had loaded, what its threads looked like, and what version of Mac OS I'm running.
tl;dr: Get your facts straight, what they're doing isn't perfect but it's a hell of a stretch to call it spying or even restrictive to all but the dumbest users.
On a related note I wonder what the dollar amount of broadcast standard related patent royalties is rolled into the typical price of an HDTV vs. an equally resolution equipped computer monitor.
The only part I can find solid information on is the MPEG license, which is $2 per unit at this time (previously $4, then $2.50). Some claim the total cost can be up to $30 per unit, but that number is from a group fighting the licensing cost so take it with appropriate quantities of salt.
A new international television standard. How long until we in the US invent our own entirely incompatible system just so it can depend on patents owned by American companies?
ATSC versus DVB-T, CDMA2000/EvDO vs. GSM/UMTS, etc.
*Your* firewall. The one that you control on/your/ end. Not the remote user's firewall.
Yes their end. If the user performing the remote support is connected via a mobile phone, they may not be able to receive incoming connections.
I've had to do it many times, I'm out on the road and something comes up for which I'm either the only person that can handle it or the best person to handle it. I don't know what the restrictions will be on whatever internet connection I'm attached to, but if I can make outbound connections I can probably get one of the services the OP listed as wanting to emulate working. That's the goal here, neither the supporter nor supportee should require the ability to receive arbitrary incoming connections at their current location.
You didn't understand the question. The question was about servers hosting multiple domains, assumedly in the context of HTTP since most other protocols don't give a fuck about the domain name. To test this properly, you'll need to either edit the HTTP request by hand or convince your machine that so and so server is actually the host you're requesting. The HOSTS file provides a convenient way to do this for those without direct control over their DNS server.
That said, unless your site is in the list of protected domains this is entirely irrelevant, and if it is you are probably running your own internal DNS which allows for as much testing as you'd like.
The sites affected are regularly accessed domains for which malware has historically been known to attack via the HOSTS file. The few users who legitimately need to add these domains to said files can be assumed to be able to figure out how to disable said restriction (though I agree with the idea that MS should have put a note in the file stating that such a thing was occurring) or run their own DNS making this a non-issue.
tl;dr: You interpreted the question wrong, but the question was pointless to begin with.
But it also does this for Doubleclick, which sounds more like someone sucking up to their corporate partners.
You do realize who owns DoubleClick, right? Google. Not exactly a partner of Microsoft. Microsoft has their own ad network that competes with DoubleClick, so that part actually helps make a case to me that this was not ill-intentioned.
It seems to make sense. Inject your own ads in place of one of the most popular ad networks. Any other content you want to bundle along with those ads you can as well of course.
This is correct. In a managed environment it's not exactly rocket science to put your cert on the computer, allowing you to resign anything HTTPS. Make it clear to the users that EVERYTHING is being monitored and they have no expectation of privacy on said computers and go for it.
Using a bogus cert that throws warnings in the browser is just an idiotic way to train your users that clicking through SSL warnings is normal.
You need to look at better ROMs and ignore the crap.
What I've observed from having a few Android devices over the last few years is that in general stock-based ROMs are garbage. Each device typically has one or two developers who actually care about making a clean, functional ROM while sticking to the OEM-provided software where possible for stability reasons. Fresh is a good example of this on the HTC side of things. Unfortunately the vast majority of stock-based ROMs are made by kids who think l33tsp34k is cool and that everyone will love their abortion of a theme. You can usually tell these apart by how much their XDA threads look like Geocities or Myspace pages, full of animated GIFs and eye-raping colors. The "developers" usually don't take well to criticism, particularly if it's about their terrible visual choices, and XDA unfortunately encourages these idiots.
AOSP ROMs such as CyanogenMod tend to be a million times better. Don't confuse this with "kang" versions thrown together by someone who is usually about one step above the aforementioned eye rapists, I'm talking about official builds and unofficial releases from people who actually work on the project. AOKP would be another similar large AOSP-based project I'd put on the same level. Their documentation tends to be terse but clear and the people working on the project generally aren't three year olds.
MIUI is another AOSP fork worth mentioning on its own due to how far they've taken the UI. It's AOSP (I think derived from CM) at its core, but it's almost like a carrier skin on top, just not shitty and designed to work on many devices and keep up with AOSP in a reasonable timeframe. This started as some bored people translating a Chinese ROM that was a blatant iOS knockoff, but has morphed in to a much larger project.
Disclaimer: My employers at the time had a much better OS, (with a high level of compatibility with Intel's ISIS (I know because I dis-assembled virtually the whole of ISIS to achieve this)) but failed to even try to sell it "Because no one would buy a British OS!"
I think Lucas electrics pretty much ruined the rest of the world's perception of anything you guys do that remotely involves electricity.
I apparently dropped a few sentences somehow. I'm aware of flash carts, I have one for my DS. I'm aiming to have something that will preferably feed from a network server, or if that's not feasible a USB connection. If I add a new game to my collection or feel like playing with another ROM hack I want to be able to just throw it in a share and have it magically end up available on the console. That's how my HTPCs and one special Xbox 360 work already, so I'm sort of spoiled by the convenience and want it everywhere.
On the other hand, all you really need for even that is the console itself, since you can get rewritable cartridges to load ROMs onto. If I were putting together an original-console setup, I'd probably do it that way, archiving the consoles physically but the games electronically.
That's a project I've been investigating the feasibility of recently, using a microcontroller or FPGA to emulate cartridges for use with real console hardware. Most ROM-only cartridges could probably be handled by a microcontroller with sufficient memory, but the 16 bit era in particular brought a lot of on-cartridge coprocessors (SNES's SuperFX being the most well known) which would probably bump those firmly up in to FPGA territory.
The catch of course is that the consoles that would be easiest to do this with are also those that are best emulated. I'm pretty sure there were recent articles about a SNES emulator (maybe simulator in this case) which on paper should be literally perfect, though it has much higher CPU requirements than others due to its exacting simulation of the individual chips rather than taking shortcuts where available.
Notice the cities I chose. They're all cities that have or are the nearest major city to massive internet exchange points. They have plenty of backbone available, the lacking portion is 100% in the last mile.
Your choice of Las Vegas as an example point is amusing, since it's one of the places seeing a boom in connectivity and datacenters due to being a major city in a place with little potential for natural disasters.
That argument would work if places that matched density with European or Asian cities also matched or approached their internet connectivity. They don't, however, not by a long shot.
Sure, someone living out in Nowhere, Idaho can't expect readily available and inexpensive broadband, but someone living in or around NYC, LA, or DC should. They don't have shit worth comparing either, for the most part. Lucky pockets of population have FTTP services or cable carriers who don't suck, but the vast majority have yet another overpriced Time Warner or guaranteed to be shit DSL.
If the Europeans can deploy these nice networks in cities that were never built to be friendly to modern infrastructure, why can't we seem to figure it out even in new construction?
Same. I ran a Windows client/Linux server environment for years with Samba 3 acting as a NT4 DC and the only real pain of managing printers was figuring out what parts I needed to strip out from HP's shitty installer to get driver auto-install working when a user first connected.
The two things I really missed from a native Windows server environment were WSUS and the ability to push software without resorting to silly hacks running in the login script. I can't wait for Samba 4 to get stable enough for packages to exist so I can play with it on my home network.
Stability problems I can deal with, compiling every time there's an update I can't stand.
And which one of those is a standalone browser? Here's a hint: Not a single one.
Dolphin is a different UI wrapped around the Safari/Webkit core. It's exactly equivalent to the Maxthon and Neoplanet IE-based browsers from the pre-Firefox dark ages. These are fully allowed and have been since day one.
Opera Mini looks like a web browser, but it's actually a renderer for Opera Binary Markup Language. Opera's servers run the real web browsers and translate page content to a mobile-friendly format intended for use on underpowered J2ME phones and old WinMo/Symbian devices where a full modern browser is not feasible. They've also released it for iOS so they can offer something there. It's the closest to a real alternative web browser as exists on the official store, but it isn't really one when you look at it. Try accessing a local server over WiFi and see how much of a web browser it actually is.
Firefox Home isn't even something that looks like a browser. It's nothing but a launcher that syncs up with your Firefox Sync account and provides access to your bookmarks and lists of open tabs from Firefox sessions. Click one though and it opens Safari.
Opera Mini is not a web browser. That's why it's allowed.
Back in the shitty J2ME phone days, Opera wanted to release a more fully featured browser for phones, but the hardware just wasn't there. What they did to solve that is create a translator that runs on their servers, optimizes the page for ultra-small screens and the features they chose to support, and sends over instructions in a binary format to a light-weight renderer frontend running on the phone itself. This is Opera Mini. For all intents and purposes it's a control interface for a remote web browser.
Versions up to 4 were released for J2ME devices, 5 and beyond have been for smartphones, aimed users on older devices that can't run full Opera Mobile properly and/or those wanting to conserve limited data (the binary format is much smaller than HTML and I think there's image compression in play as well). Somewhere along the lines the Opera folks realized that this would skirt Apple's rules and gave it a shot, the rest is history.
There are no actual browsers on the App Store (as in something that could access pages hosted on a server in my living room from my WiFi without any data leaving my network), just Opera Mini and a variety of different frontends piled on top of plain old Safari/Webkit.
and which are more tolerant of over-voltage than delicate electronic devices.
These days I'm not so sure. The bottom-of-the-line models and the units marketed to apartments and such are still mostly electromechanical, but most major appliances marketed directly at the consumer now seem to have electronic controls of some sort, even if it's little more than a clock for scheduling things.
I had to buy a washer and dryer about 18 months ago and anything above the $250 point for either of those seemed to have a digital display and some sort of keypad for setting entry.
The basic unit itself will survive as the core parts are still the same simple and durable things as before, but when the only way to turn it on is the now-fried digital control system you're just as stuck.
Slashdot Mirror
Who doesn't shut steam down after they are done playing?
People who have the vast majority of their games from the last x years in Steam and don't mind the few dozen MB of RAM it takes up in the background for the time savings of just being able to launch any game in two clicks?
People who use the Steam IM system to see when their friends are online and playing a game they might want to join?
And who doesn't close all programs before shutting down?
Who manually closes any of their autorun-at-startup programs before shutting down? Or really anything? Make sure anything open you care about is saved, Start, Shutdown.
What the hell yourself? Are you one of those crazy people who thinks empty RAM is a good thing? Before I had a SSD the Steam client took about 15 seconds to reach the login screen, then another 20-30 before my game list was available. Now these are in the 5 and 10-15 second ranges respectively, but still more than just right-clicking the Steam icon in my tray or pinned to my taskbar and going right to the game I want. Unless you're really tight on RAM or want to prevent people from playing your Steam-purchased games on your computer without logging in first, just treating it as a constantly running service is a good common choice.
There was a long-standing bug in Steam which made offline mode unreliable for anyone who shuts their computer down regularly.
When you'd shut down Windows, Steam would apparently just sit around and fail to heed the warnings the OS gave it until it would be forced to close. This made it not properly save the offline mode token, so if the next time you launched the client was without an internet connection, you were SOL.
If you closed Steam before shutting down it worked fine, but since this wasn't common knowledge and you don't usually plan internet outages at the consumer level as far as an average user cares it didn't work.
The actual cause of the bug was identified publicly by a user a few months ago and shortly followed by a Steam client update which resolved the problem. At this point offline mode works exactly as expected.
They do have a torrent, just they hosted the tracker on GameUpdates which pretty much crumbled under the weight. Why they didn't use one of the many public open high-capacity trackers I do not know.
My company had a need for a file server. I was the most technically knowledgeable in the company, so the task fell on me. It was being done on a zero budget using a spare computer, so I chose Linux. The rest you can guess from there.
You mean the OS that, by default, blocks you from running content that isn't blessed by Apple? Yes, you can download apps from sources that aren't the App Store - but they still have to be signed, otherwise, it either will refuse to run or lie to you and say that the app is "damaged" and you should "drag it to the trash."
Try not to make shit up, it makes your argument look like the rant of a crazy fanboy to anyone who knows better regardless of whether other parts are valid.
First, it's not anything about content blessed by Apple. Any registered Apple developer gets a key with which they can sign all their apps as they desire. Apple approval only matters for their App Store. Gatekeeper can restrict a computer to App Store apps only, but that is not the default setting and must be chosen by the user.
Admittedly the developer registration costs $99/year to maintain. This should not be a problem for any commercial developer, even at the smallest scale, but obviously may be a barrier for freeware developers. Anyone developing for iOS (officially) is already a member as well, making it effectively free for them. I would note as well that a key for Microsoft's comparable Authenticode system costs three times as much per year, though it is not currently enforced by the operating system in many user-visible ways (the most notable one being the dialog before launching downloaded EXEs).
Second, where the hell are you getting this idea that it would lie and say the application is damaged? That happens if the application IS signed but does not match what's expected based on the signature. An unsigned app in the default configuration gets a dialog that says it's been blocked by the computer's current security policy without in any way implying that such apps are necessarily bad. The dialog practically comes out and tells the user where to change this as well, the few who couldn't figure it out from that point are exactly the kind of user this is intended to protect.
When the Gatekeeper feature is switched off (I haven't upgraded to 10.8 yet so I can't confirm or deny your claim of a warning of dire consequences while doing this), the warning on unsigned applications changes to one which is practically identical to the one Windows displays on downloaded apps. It only shows once, then goes away forever for that application.
Third, sending diagnostic and usage data is disabled by default. It asks to send such data after certain kinds of failures, and if you select "Don't Ask Me Again" before clicking yes then sure it will enable itself, but that was your choice as the user. The data collected (and sent if enabled) is available to the user from the "Console" app, same as any other logs on the system. Remember, these are still computers on which the user can run any apps they please, inspect the hard drive as they please, and change any certificates they please. What's sent by the computer can be inspected by the user with ease. If there was a difference between what Apple claims, what Console shows, and what's actually being logged and transmitted it'd be trivial to see and call them out on it.
The only time Apple knows what non-AppStore apps you have on your Mac is if there's a crash which you choose to report to Apple (or have previously chosen to send all to them) and the app is actually running at the time of the crash. Even in these cases it's not guaranteed, as the OS can decide that other apps weren't relevant to the problem in which case it doesn't bother logging them. I just looked at the five crash logs I have since installing the OS, and not a single one has even a process list in it. All that is logged is the application that crashed, how it crashed, information about its memory usage, what libraries it had loaded, what its threads looked like, and what version of Mac OS I'm running.
tl;dr: Get your facts straight, what they're doing isn't perfect but it's a hell of a stretch to call it spying or even restrictive to all but the dumbest users.
On a related note I wonder what the dollar amount of broadcast standard related patent royalties is rolled into the typical price of an HDTV vs. an equally resolution equipped computer monitor.
The only part I can find solid information on is the MPEG license, which is $2 per unit at this time (previously $4, then $2.50). Some claim the total cost can be up to $30 per unit, but that number is from a group fighting the licensing cost so take it with appropriate quantities of salt.
A new international television standard. How long until we in the US invent our own entirely incompatible system just so it can depend on patents owned by American companies?
ATSC versus DVB-T, CDMA2000/EvDO vs. GSM/UMTS, etc.
*Your* firewall. The one that you control on /your/ end. Not the remote user's firewall.
Yes their end. If the user performing the remote support is connected via a mobile phone, they may not be able to receive incoming connections.
I've had to do it many times, I'm out on the road and something comes up for which I'm either the only person that can handle it or the best person to handle it. I don't know what the restrictions will be on whatever internet connection I'm attached to, but if I can make outbound connections I can probably get one of the services the OP listed as wanting to emulate working. That's the goal here, neither the supporter nor supportee should require the ability to receive arbitrary incoming connections at their current location.
You didn't understand the question. The question was about servers hosting multiple domains, assumedly in the context of HTTP since most other protocols don't give a fuck about the domain name. To test this properly, you'll need to either edit the HTTP request by hand or convince your machine that so and so server is actually the host you're requesting. The HOSTS file provides a convenient way to do this for those without direct control over their DNS server.
That said, unless your site is in the list of protected domains this is entirely irrelevant, and if it is you are probably running your own internal DNS which allows for as much testing as you'd like.
The sites affected are regularly accessed domains for which malware has historically been known to attack via the HOSTS file. The few users who legitimately need to add these domains to said files can be assumed to be able to figure out how to disable said restriction (though I agree with the idea that MS should have put a note in the file stating that such a thing was occurring) or run their own DNS making this a non-issue.
tl;dr: You interpreted the question wrong, but the question was pointless to begin with.
But it also does this for Doubleclick, which sounds more like someone sucking up to their corporate partners.
You do realize who owns DoubleClick, right? Google. Not exactly a partner of Microsoft. Microsoft has their own ad network that competes with DoubleClick, so that part actually helps make a case to me that this was not ill-intentioned.
It seems to make sense. Inject your own ads in place of one of the most popular ad networks. Any other content you want to bundle along with those ads you can as well of course.
This is correct. In a managed environment it's not exactly rocket science to put your cert on the computer, allowing you to resign anything HTTPS. Make it clear to the users that EVERYTHING is being monitored and they have no expectation of privacy on said computers and go for it.
Using a bogus cert that throws warnings in the browser is just an idiotic way to train your users that clicking through SSL warnings is normal.
You need to look at better ROMs and ignore the crap.
What I've observed from having a few Android devices over the last few years is that in general stock-based ROMs are garbage. Each device typically has one or two developers who actually care about making a clean, functional ROM while sticking to the OEM-provided software where possible for stability reasons. Fresh is a good example of this on the HTC side of things. Unfortunately the vast majority of stock-based ROMs are made by kids who think l33tsp34k is cool and that everyone will love their abortion of a theme. You can usually tell these apart by how much their XDA threads look like Geocities or Myspace pages, full of animated GIFs and eye-raping colors. The "developers" usually don't take well to criticism, particularly if it's about their terrible visual choices, and XDA unfortunately encourages these idiots.
AOSP ROMs such as CyanogenMod tend to be a million times better. Don't confuse this with "kang" versions thrown together by someone who is usually about one step above the aforementioned eye rapists, I'm talking about official builds and unofficial releases from people who actually work on the project. AOKP would be another similar large AOSP-based project I'd put on the same level. Their documentation tends to be terse but clear and the people working on the project generally aren't three year olds.
MIUI is another AOSP fork worth mentioning on its own due to how far they've taken the UI. It's AOSP (I think derived from CM) at its core, but it's almost like a carrier skin on top, just not shitty and designed to work on many devices and keep up with AOSP in a reasonable timeframe. This started as some bored people translating a Chinese ROM that was a blatant iOS knockoff, but has morphed in to a much larger project.
Disclaimer: My employers at the time had a much better OS, (with a high level of compatibility with Intel's ISIS (I know because I dis-assembled virtually the whole of ISIS to achieve this)) but failed to even try to sell it "Because no one would buy a British OS!"
I think Lucas electrics pretty much ruined the rest of the world's perception of anything you guys do that remotely involves electricity.
I dropped a few sentences.
I dropped a few sentences.
I apparently dropped a few sentences somehow. I'm aware of flash carts, I have one for my DS. I'm aiming to have something that will preferably feed from a network server, or if that's not feasible a USB connection. If I add a new game to my collection or feel like playing with another ROM hack I want to be able to just throw it in a share and have it magically end up available on the console. That's how my HTPCs and one special Xbox 360 work already, so I'm sort of spoiled by the convenience and want it everywhere.
On the other hand, all you really need for even that is the console itself, since you can get rewritable cartridges to load ROMs onto. If I were putting together an original-console setup, I'd probably do it that way, archiving the consoles physically but the games electronically.
That's a project I've been investigating the feasibility of recently, using a microcontroller or FPGA to emulate cartridges for use with real console hardware. Most ROM-only cartridges could probably be handled by a microcontroller with sufficient memory, but the 16 bit era in particular brought a lot of on-cartridge coprocessors (SNES's SuperFX being the most well known) which would probably bump those firmly up in to FPGA territory.
The catch of course is that the consoles that would be easiest to do this with are also those that are best emulated. I'm pretty sure there were recent articles about a SNES emulator (maybe simulator in this case) which on paper should be literally perfect, though it has much higher CPU requirements than others due to its exacting simulation of the individual chips rather than taking shortcuts where available.
Notice the cities I chose. They're all cities that have or are the nearest major city to massive internet exchange points. They have plenty of backbone available, the lacking portion is 100% in the last mile.
Your choice of Las Vegas as an example point is amusing, since it's one of the places seeing a boom in connectivity and datacenters due to being a major city in a place with little potential for natural disasters.
That argument would work if places that matched density with European or Asian cities also matched or approached their internet connectivity. They don't, however, not by a long shot.
Sure, someone living out in Nowhere, Idaho can't expect readily available and inexpensive broadband, but someone living in or around NYC, LA, or DC should. They don't have shit worth comparing either, for the most part. Lucky pockets of population have FTTP services or cable carriers who don't suck, but the vast majority have yet another overpriced Time Warner or guaranteed to be shit DSL.
If the Europeans can deploy these nice networks in cities that were never built to be friendly to modern infrastructure, why can't we seem to figure it out even in new construction?
Same. I ran a Windows client/Linux server environment for years with Samba 3 acting as a NT4 DC and the only real pain of managing printers was figuring out what parts I needed to strip out from HP's shitty installer to get driver auto-install working when a user first connected.
The two things I really missed from a native Windows server environment were WSUS and the ability to push software without resorting to silly hacks running in the login script. I can't wait for Samba 4 to get stable enough for packages to exist so I can play with it on my home network.
Stability problems I can deal with, compiling every time there's an update I can't stand.
Just another Safari/Webkit frontend.
http://itunes.apple.com/us/app/firefox-home/id380366933?mt=8
http://itunes.apple.com/us/app/dolphin-browser/id452204407?mt=8
http://itunes.apple.com/us/app/opera-mini-web-browser/id363729560?mt=8
And which one of those is a standalone browser? Here's a hint: Not a single one.
Dolphin is a different UI wrapped around the Safari/Webkit core. It's exactly equivalent to the Maxthon and Neoplanet IE-based browsers from the pre-Firefox dark ages. These are fully allowed and have been since day one.
Opera Mini looks like a web browser, but it's actually a renderer for Opera Binary Markup Language. Opera's servers run the real web browsers and translate page content to a mobile-friendly format intended for use on underpowered J2ME phones and old WinMo/Symbian devices where a full modern browser is not feasible. They've also released it for iOS so they can offer something there. It's the closest to a real alternative web browser as exists on the official store, but it isn't really one when you look at it. Try accessing a local server over WiFi and see how much of a web browser it actually is.
Firefox Home isn't even something that looks like a browser. It's nothing but a launcher that syncs up with your Firefox Sync account and provides access to your bookmarks and lists of open tabs from Firefox sessions. Click one though and it opens Safari.
Opera Mini is not a web browser. That's why it's allowed.
Back in the shitty J2ME phone days, Opera wanted to release a more fully featured browser for phones, but the hardware just wasn't there. What they did to solve that is create a translator that runs on their servers, optimizes the page for ultra-small screens and the features they chose to support, and sends over instructions in a binary format to a light-weight renderer frontend running on the phone itself. This is Opera Mini. For all intents and purposes it's a control interface for a remote web browser.
Versions up to 4 were released for J2ME devices, 5 and beyond have been for smartphones, aimed users on older devices that can't run full Opera Mobile properly and/or those wanting to conserve limited data (the binary format is much smaller than HTML and I think there's image compression in play as well). Somewhere along the lines the Opera folks realized that this would skirt Apple's rules and gave it a shot, the rest is history.
There are no actual browsers on the App Store (as in something that could access pages hosted on a server in my living room from my WiFi without any data leaving my network), just Opera Mini and a variety of different frontends piled on top of plain old Safari/Webkit.
and which are more tolerant of over-voltage than delicate electronic devices.
These days I'm not so sure. The bottom-of-the-line models and the units marketed to apartments and such are still mostly electromechanical, but most major appliances marketed directly at the consumer now seem to have electronic controls of some sort, even if it's little more than a clock for scheduling things.
I had to buy a washer and dryer about 18 months ago and anything above the $250 point for either of those seemed to have a digital display and some sort of keypad for setting entry.
The basic unit itself will survive as the core parts are still the same simple and durable things as before, but when the only way to turn it on is the now-fried digital control system you're just as stuck.