> Hint: look at Amtrak. It costs significantly more to travel per mile on Amtrak than it does to fly; > it's slower, more cramped, less friendly, and you've got to wait longer.
Yes, it's more expensive to travel on Amtrak than to fly (at least, if you get a room), and it's a lot slower. However, there's no way in HELL it's more cramped than flying. Even in a roomette shared with one other person, upper bunk stowed, you have more real room than you'll ever have in first class on any domestic airline. Ditto, for seats in coach.
As for waiting... well, yeah... you might have to wait longer. But the quality of waiting for a train is different from waiting for a plane. When the train shows up, you can walk aboard without even having to shut down your laptop, or "discontinue use of approved personal electronic devices." If you're tethered to your cell phone, you might not even lose your current IP address unless you voluntarily disconnect at some point.
Amtrak is far from perfect. In fact, it's a pretty sad excuse for a passenger rail network. But even in its imperfect, sad state, it's nicer than the hellish ordeal flying has become over the past 10 years. Given a choice between a 3 hour trip in a shiny new overcrowded city bus, or a 5 hour trip in a slightly-tattered, but roomy and comfortable limo, I'll take the slightly ratty limo.
> Tampa and Orlando, a grand distance of 85 miles, or about 90 minutes driving.
On its own, you're absolutely right. A high-speed train that goes nowhere besides downtown Tampa to Orlando International Airport (with Disney sort of a bus ride detour along the way) is stupid. Where FDOT has a chance to redeem itself is if, by some miracle, it decides to build tracks suitable for "true HSR", but buys TRAINS capable of running at 125+mph between Tampa and Orlando, and equally capable of leaving those shiny new tracks near Auburndale and heading south to West Palm Beach, Fort Lauderdale, and Miami along the CSX tracks that already exist (possibly double-tracked, and improved a bit to allow 110mph operation for most of the way between Sebring and West Palm Beach, and 70-80mph the remainder of the way down to Miami). THEN, they make sense.
For the past 20+ years, FDOT's own studies have basically come to two conclusions: HSR will always hemorrhage money, and 90-110mph intermediate-speed rail will make money hand over fist. Why? The market is already here now for fast rail travel between South and Central Florida. Driving from Miami to Orlando sucks, and driving to Tampa sucks even more... but not *quite* badly enough to endure the hell and misery of flying. At ISR speeds, the corridor between Auburndale and Miami already exists; it's even double-tracked to passenger-rail standards from WPB to Miami. So, why would HSR hemorrhage cash? Because it would cost about $5-10 billion to build a really, really nice ISR passenger rail network connecting Miami, Orlando and Tampa... and $50-100 billion to build True HSR(TM) every last inch of the way down to Miami. $5-10 billion can be rationalized... it's basically $250-500 per Floridian. If nothing else, it would spur cross-state tourism like nothing the state has ever done.
And yeah, people would step off the train and grab a rental car, just like if they flew. The difference is, instead of spending 3 hours running and "actively waiting" (being forced to do nothing besides sit or stand and wait for permission to do something), they'd spend the same amount of time eating, drinking, or otherwise enjoying the first or last part of their trip. Ideally, passengers could do the rental car formalities on the train, and walk off at their destination with their rental car's keys in hand.
As for the HSR part between Tampa and Orlando, in an indirect way, it DOES still sort of make sense, even if the rest of the line is ISR. Why? Well, the tracks from Miami to WPB are already owned by FDOT and used by Tri-Rail (commuter rail). The tracks from WPB to Auburndale (halfway between Tampa & Orlando, about 5 miles south of I-4) are owned by CSX, but almost useless for freight. The tracks between Auburndale and Jacksonville all but been sold outright to FDOT (in fact, a segment has). HOWEVER... the tracks between Auburndale and Tampa are very, VERY valuable to CSX... and CSX has made it clear they won't give them up, or even share them, without a fight. And in Florida, they can fight dirty. It's right in the constitution... railroads can condemn adjacent land to their existing right of ways for basically any purpose without limit. So, if FDOT tried to use eminent domain to take the northern 50 feet of CSX's corridor, CSX will use eminent domain to take the next 50 feet or more to the south. And a few cities, including Lakeland, would be very VERY angry if FDOT induced CSX to do that. So FDOT doesn't dare.
*THAT* is why it sort of makes sense (at least, purely from the perspective of Florida's self-interest) to build the new tracks down I-4. Now, before the Feds showed up with a dumptruck full of money, FDOT's plans were a bit less ambitious... basically, a new track down I-4 to somewhere around Auburndale, then CSX the rest of the way to Orlando. But since the Feds are now eating a huge chunk of the cost, it's cheaper for Florida to just humor the feds and build 100% HSR all the way to Orlando.
NOW, IMHO, FDOT has the Orlando end all wrong. Instead of building
> and it's free- something that ought to appeal to poor starving college students.
Don't forget, if you're a college student, Microsoft software *is* practically free. Also, the "poor starving student" meme is grossly exaggerated, at least in the US, if you basically count full-time students at a major university (private OR public) who live on campus. Even if they ARE legally penniless, their cash-flow problems aren't going to kick in until after they've graduated and have to start paying back their student loans and maxed-out credit cards.
> but I do not think Linux is such a terrible operating system that it would see no use whatsoever
I think their statistics are skewed by only counting "Linux" if that's the one and only operating system on the user's main computer. Round up some students majoring in computer engineering or computer science, and I think you'd find that close to 100% of them have a small pile of laptops and/or additional partitions with Linux on them. They just aren't running Linux as their One True Daily Operating System.
> It was not patented. As such, the manufacturing will not go to China, but to Japan (who will then take it to China). > Personally, Corning has earned my disdain. At this time, I will quit buying from Corning.
Er... why? If anything, keeping it a trade secret would have been a calculated gamble based on the apparent lack of real-world applications for it 60 years ago and lack of anticipated applications for it within the next 20 years.
Besides, even without patents and trade secret law, there's trademark law. At the end of the day, there will inevitably be factories in China making similar glass... but we all know that none of them will ever duplicate it exactly, even if they could, because the temptation will always be there to cut corners in some way people are unlikely to notice until it's too late. All Corning has to do is wait for someone to use the name of Gorrilla Glass in vain, and unleash the lawyers on them.
Could companies in Europe and Japan duplicate it? Sure. But they won't, because their manufacturing costs are as high (or higher) than Corning's, and they don't have the benefit of Corning's additional research on how to manufacture it cost-effectively. By the time they had it ready to sell, they wouldn't be particularly cost-competitive anyway. Like it or not, Corning pretty much owns the specific market segment for "expensive, exceptionally high-quality specialty glass for niche uses".
> Those places, there is actual broadcast TV that people can easily watch (without being dependent on the data bandwidth.)
Part of the problem is that American HDTV uses 8VSB instead of COFDM for data transmission. Oversimplifying a bit, 8VSB is great if you want the longest range over flat terrain possible AND assume your audience has properly-grounded directional antennas... but it's difficult to receive in a moving vehicle (even WITH a good antenna), and nearly impossible to receive by a pocket receiver with stub or internal antenna. In contrast, COFDM has a lot less "fringe" range, but has fairly minimal antenna requirements if you're within a few miles of the tower.
The net result is that in Europe and Japan, a pocket-sized HDTV tuner is likely to work well in urban areas. In America, you'd be lucky to get a viewable HDTV signal anywhere besides maybe Weehawken (New Jersey... high on a cliff, line of sight to the antennas on the Empire State Building ~3 miles away).
In other words, it was one of those, "Oops" moments when a new standard got declared by law, and nobody realized the oversight until it was too late to do anything about it. Or so the TV industry claimed, right before rendering every HDTV sold up to that point effectively obsolete for its intended purpose by virtue of not supporting HDCP...
> It's most telling that Palm is flatlining and Windows Mobile has lost half of its already > meager market share in the past year.
Are you counting people who own phones that were sold with Windows Mobile, but are now running Android (like the HTC Touch HD2)? The HD2 debacle will go down in tech history as one of Microsoft's worst marketing/business decisions in history. Here's a phone that was eagerly embraced by Microsoft's few remaining enthusiasts, even as their friends and peers ran for the door marked "Android", only to get its owners metaphorically kicked in the balls by Microsoft on what was probably the lamest pretense for non-compatibility *ever* (it had four buttons instead of three).
Microsoft could hardly have done a better job of driving its few remaining friends into the Android camp if they'd personally rebranded MSDN as an Android portal & given a free Nexus One to everybody who attended a Microsoft event in 2010.
My apologies. I'd read a few summaries earlier in the day, and was under the impression that it was up to the app itself to persistently keep track of the app's licensing state if it didn't want to check every single time.
Now, if only Google could either bully HTC, Samsung, Motorola, and the rest into leaving HID enabled when they build the kernel (so AOSP could implement the remainder at some later point without needing a new kernel), or come up with some good way to allow users with 100% compiled-from-the-ground-up (kernel and all) AOSP builds to run protected Market apps. Say, by distributing the DRM support as a loadable kernel module augmented by the DRM built (by definition, since that's what officially differentiates "SD" from "MMC") into every (micro)SD card ever sold.
There were (at least) two fundamental flaws with the original Android Market protection scheme, neither of which appears to have been rectified by this change (besides possibly to make matters worse for end users):
* As everyone has already noted, lots of people around the world with Android phones can't actually buy apps from Android Market, EVEN IF they have a Mastercard/Visa/AMEX card with dollar-denominated account. That's just plain fucked.
* You can't officially purchase and run protected Market apps if your phone is running an unblessed "Developer" kernel. Of course, there's not a single goddamn phone from HTC, Samsung, or Motorola with Google-blessed kernel that has BlueZ Bluetooth HID profile compiled into it, so it's impossible to build your own kernel with it enabled without being formally exiled from 99% of commercial Android apps. At least, unless you crack them. Any DRM scheme that forces legitimate users to crack apps they purchased in order to use them is fundamentally broken, especially when there are still gaping holes in Android phones that need a customer kernel to fix.
As for "developer's option" whether or not to cache, let's be honest... at least half the developers publishing commercial apps don't have the slightest clue in HELL how to implement a secure caching scheme, and they aren't going to purchase a proprietary one that demands more money up front than they're likely to earn from the app's sale. So, anybody care to guess what's going to happen? Most apps in Market are going to end up checking the server every goddamn time, because the alternatives are too hard/expensive for most Android publishers to deal with. IMHO, Google got THAT part EGREGIOUSLY wrong. They should have distributed the Android DRM module themselves, and made it free & easy for publishers to do cached checking, but left it difficult and minimally-documented how to bypass that caching and check the server every time.
I love Android. I really do. But it's so incredibly frustrating when Google turns around and fucks things up in ways that CAN'T be fixed by end users with access to Android's sourcecode... usually, mistakes that are almost incomprehensible given the amount of in-house talent and expertise Google has available to it. At times, Google actually manages to make even *Microsoft* look coherent and customer-focused.
> Also didn't we have all those things about 100 years ago?
Exactly. If anything, it could almost be argued that the pollution in late 19th-century Britain, France, and Germany (and parts of America, for that matter) were noxious/toxic enough to make the most badly-polluted square mile of China look like the Garden of Eden by comparison. At least people in China don't have to rely on wood and coal-burning stoves & fireplaces for cooking and heating ON TOP of the pollution being produced by factories (at least, urban factory workers who live amidst the worst pollution) don't.
As a species, humans are easy to kill individually, but surprisingly difficult to effectively exterminate. The dinosaurs didn't have preserved food, hydroponics, artificial lighting, and global distribution networks, so when the skies went dark and 99% of photosynthesis shut down for a few years after the impact event, they were screwed. A similar event would be an unprecedented human tragedy, but the likelihood of enough humans surviving to repopulate the Earth eventually is practically assured.
Someone please mod this up. The games in danger of fading into oblivion aren't Pac Man, Dig Dug, and Centipede... or even the more obscure games from that era, or even the 90s. The games in danger are the DRM'ed MMORGs that have no existence independently of the company hosting them online.
Historically, most preservation is done by people who aren't the official owners. Just look at Hollywood -- execs used old, only-existing-copies of silent films from the 20s to start bonfires at beach parties for years. Most of the old films we have copies of came from private collectors, many of whose copies were technically (if not outright) illegal for them to own.
It's one reason why librarians are more than slightly disturbed about predictions that ebooks will replace bound, printed volumes. It's not that they're anti-technology... it's the fact that books have a way of persisting and remaining accessible long after whomever owns the copyright has lost interest in them for any purpose besides random windfalls due to infringement lawsuits. In contrast, DRM-protected content can go away for reasons ranging from active neglect all the way to intentional retraction. It's pretty much a given that cracking 1024-bit encryption will be do-able 50-100 years from now, but the danger with digital content is the combination of its short media life and persistence of copy protection. If ebook readers protect content not only by encryption, but by copy-protection as well, and the media doesn't retain its integrity long enough for the DRM itself to be crackable... well, the problem is obvious.
I think it's not so much a matter of not knowing about this as a potential vulnerability, as it is a case of the hardware necessary to pull it off suddenly becoming cheap and affordable to just about anyone with the slightest interest in doing it.
Perfect illustration of "exploit" that becomes possible due mainly to falling prices:
My best friend owns two Blu-Ray players. One is Region "A". He bought it for $99 the day after Thanksgiving last year. The other is Region "B". He paid around $160 for it, including shipping, from the guy in Hungary or Romania who was selling them on eBay. Both are perfectly happy to feed 1080p24 video via hdcp-protected hdmi to his TV. If he cared, I'm sure he could buy a Chinese Region "C" Blu-Ray player for another hundred bucks or so, plus maybe another $50 for a HDMI switcher. Region coding only works as long as players are too expensive to just go out and buy one from every region you care about.
It depends. From what I remember (teen in the late 80s), at least in the Apple//e, Atari 400/800, Vic20, TI99/4A, C64, Atari ST, and Amiga era, most games didn't start out as a brilliant high-level concept. Someone discovered a cool video hack, and managed to build a game around it. That was part of the reason why ported games on any platform almost universally sucked... whatever it was that made them look GOOD on their native platform didn't exist (or was moot and no big deal on the new platform), and forced to stand on their own, they fell flat on their face.
The games that lasted were the ones that did, in fact, transcend the video hack that made them possible in their first implementation. Jumpman is still fun, warts and all. In contrast, most of us wouldn't waste the erase cycle on a flash drive for a pirated copy of Incredible Mission, because the only thing that ever made it interesting was the relatively high-resolution (for the time) graphics and 4 seconds of digitized speech at the start. A hundred years from now, at least a few people will still know what Zork was. A few might even have played it in some context. Nobody will care about playing Elvira, Mistress of the Dark on an Amiga emulator.
Actually, you wouldn't even need a vector graphics monitor. Given a decent 20" glossy 1920x1080 LCD (possibly in portrait mode) capable of 120hz native updates with subpixel control via hdmi or dvi, combined with triple-buffered video emulating the bloom, defocusing, and (in the case of a color vector display like the ones used for Defender and Tempest) fringe artifacts caused by misaligned shadow masks, you could emulate a vector display with more or less perfect accuracy.
Emulated games looked like crap on 1024x768 LCD panels because they didn't scale well, and didn't offer enough raw control over the rendered video to recreate the appearance of an interlaced CRT. When you're talking about a display that can do 4x oversampling with respect to both resolution and framerate, and a computer fast enough to emulate the phosphor behavior of a CRT, you can achieve nearly perfect emulation. Hell, if you made a Vectrex-sized LCD with the pixel density of a Retina display, you could probably emulate a monochrome Vectrex down to the exact bluish-white raster on a grayish background. We're not *quite* at *that* point yet, but in another 5-10 years, we absolutely will be. At that point, having physical vector displays ceases to matter for the experience of reliving the past, and really matter only to act as a reality-check against creeping "improvements" beyond that point to the emulated video algorithm that make it look better, but destroy its historical accuracy. Does anybody *really* want to emulate the appearance of burnt-in phosphors, at least while actually *playing* an old game?
Let's face it... most 3-4 year old games at mall video arcades looked like shit thanks to years of burn-in, bloom, and slowly-dying electrolytic capacitors on the circuit board (not dying as quickly as "turn of the century" bad caps, but still visibly degraded compared to when they were new).
> Because the actual GLP'd kernel code is available, just without the proprietary drivers. > There is no violation because the source code is available.
Actually, no. It *is* pretty much a clear-cut open and shut GPL2 violation. FSF can argue with Linus over whether or not loadable kernel modules are or are not part of the kernel proper (and thus subject to requirements that their source be released), but I don't think there's *anyone* who's going to argue that what HTC does is OK.
The problem is, you can't just go and sue someone for "violating the GPL(2)". You have to prove in court that:
1) You have standing to bring the case (ie, you're one of the people who collectively own the Linux kernel's copyright)
2) The court you've chosen is the proper venue to pursue the case.
3) The code you contributed is in the kernel they shipped.
4) The source files they released were legally inadequate to fulfill their obligations as a licensee under the GPL2
5) You suffered real harm due to their actions.
Getting past step 1 could easily cost tens of thousands of dollars and involve multiple court appearances. Rest assured, the defendant's law firm is going to do everything they can to cast doubt upon your standing. If that fails, they're going to do everything they can to challenge your choice of venue (ie, the authority of the court to hear your case, and its appropriateness).
3? The easy part. Don't smile yet, because 4's going to be a bitch.
4) Have fun proving they violated the GPL2. Common sense might dictate it, but there's surprisingly little case law to actually cite one way or another because most lawsuits involving the GPL end up getting settled at the last minute & vanish from the legal radar.
5) This is the toughest of all. To get the grand prize you really want -- equitable relief granting a plea to force them to "go forth and sin no more", you have to prove that their actions have harmed you. The best-case here is probably if you own the device they shipped without the source and was unable to build your own copy of the kernel you helped develop and partially own because of their infringement. Of course, if you had to root your phone to get it into a state where it's physically possible to make use of such a kernel, you can bet they're going to throw every legal theory they can at you in the hope something will stick and get you classified as having "unclean hands" (ie, you're at least partly responsible for your plight). They might prevail, they might not, but they'll fight hard & fight dirty.
However, it doesn't end there. Suppose the judge agrees that they were totally wrong, harmed you & the larger community, and agrees to issue a court order demanding that HTC release the source to everything included in their kernel. You can bet that before anyone at HTC fires up a text editor to go to work on preparing the source, Qualcomm and everyone else who furnished those proprietary binaries (or info under NDA necessary to implement them) will have injunctions of their own to stop HTC from releasing source they don't have the right to release. If you're lucky, you might be able to force HTC to do what they should have done in the first place: re-implement them as proper loadable kernel modules, rebuild the entire kernel so it works with them, and release THAT... and do the same for everything going forward. Realistically, the likelihood of this happening fewer than 5 years from the moment you walked into the law firm's office to kick off the case is depressingly slim.
Oh... also... if the trial DOES drag on for years... don't let HTC's private investigator catch you using a different phone. Courts won't hear cases that are moot. If your claim that they've caused real harm to you rests upon being unable to build a kernel of your own for your phone (based on their infringing release), and they can demonstrate you haven't touched that phone in 3 years... well... let's just say it wouldn't be good. It wouldn't necessar
I can't speak for the Droid, but the specific problem lots of people have building kernels for HTC phones is the fact that HTC compiles their proprietary binary drivers straight into the kernel itself, then distributes the whole thing as a big monolithic blob. When it's time to distribute the source, they go in, rip out the source to anything proprietary, and dump the rest into a tarball for download. The problem is, you need the functionality provided by those proprietary binary drivers to do use things like wi-fi and the camera, but because they're effectively raw, undocumented binary blobs, you can't easily build a new kernel that incorporates them into it. Worse, all of the furnished build scripts released by HTC fail, because they attempt to reference 'include' files that aren't... well... included.
The PROPER, Linus-blessed way to incorporate proprietary binary drivers into a Linux kernel is to distribute them as loadable kernel modules. By defining a clear public interface and drawing a clean line between what's proprietary and what's not, you can then create a buildfile that compiles the part you DO have the source to in a way that works seamlessly with the binary kernel modules to which you don't, and everything "Just Works". FSF can bitch about not having the source to the WLAN drivers, but if the choice is binary WLAN drivers or no WLAN at all,.ko loadable kernel modules are a pragmatic compromise that lets you keep the functionality provided by them without breaking your ability to rebuild the rest of the kernel yourself.
> You'd think the public would want to dash towards free calls.
Because, in America at least, there's no such thing as free/low-cost wireless service that's pervasive AND works with anything vaguely resembling a high-end Android phone. American wireless service is kind of like a nightclub with a high cover charge, but free drinks and all you can eat buffet once you're past the doorman & cashier. If you're paying Sprint $90/month for "unlimited everything", there's no incentive to screw with Skype unless you have to make international calls. And with most calling cards seemingly hovering at around 2.9c/minute for international calls, even those aren't worth the hassle and latency unless you're wearing a headset and watching synchronized DVDs together with someone on the other side of the world for hours at a time.
> You will never get modded up as you deserve. I think this has been the clearest description of the hypocritical > thinking Android fans (notice didn't say fanbois...awful term) have vs. iPhone.
The difference is, we openly and routinely excoriate HTC, Motorola, and the others, bitch about their GPL violations to anyone who'll listen, and compare them to Satan's lovechild for even the smallest transgression. iPhone fans rationalize and justify Apple's behavior, and act like it's somehow shameful to demand full control of your phone.
Android owners bitch about the difficulty of building an Android distro from scratch without the cooperation of the phone's maker, usuall caused by things like HTC shitting monolithic binary kernel blobs on the curb, sniffing them a few times, and walking away satisfied instead of building their proprietary binary kernel drivers as loadable kernel modules the way they're REQUIRED to under the GPL (so new kernels can be built around them without losing the functionality provided by the.ko modules themselves).
It's not hard to imagine a few thousand angry Android users staging a protest in the Googleplex parking lot over some perceived betrayal of Android's open ideals. Try to imagine even a few dozen iPhone owners picketing on the sidewalk in front of Apple if AT&T somehow managed to push out an update that revoked root and forcibly reflashed a million jailbroken iPhones. It's almost inconceivable. Even if there were a few dozen angry iPhone owners, they'd be drowned out by the ocean of Normal Users(tm) bleating about how they shouldn't have jailbroken their iPhones in the first place, because jailbroken iPhones makes Steve Jobs sad.
> Hell, on a drive from Houston to Dallas or Austin I can tether the thing to my laptop and stream Grooveshark > with barely missing a beat, and that is a bit of a rural area between here and there.
T-Mobile, like Sprint, Verizon, AT&T, and everyone else, deploys coverage in rural areas first to highway corridors between their "real" coverage areas. It's a lot cheaper to throw a tower every couple of miles next to a major interstate along the "Texas Tee" than it is to keep building those towers further away from the road. I can drive from Fort Lauderdale to Naples along Alligator Alley and enjoy EVDO the entire way... but if I pulled off the road at one of the rest stops, got into a canoe, and went two or three miles north or south, I'd be lucky to have voice service that worked.
You can see this pretty acutely if you take Amtrak from South Florida to Orlando, and compare coverage along the way to the coverage you have while driving along the Turnpike. On the Turnpike, you'll have 3G coverage from pretty much everyone -- Sprint, Verizon, T-Mobile, and AT&T -- all the way to Orlando. It might drop to 2.5G for a few miles ~50 miles southeast of I-4, but for the most part it's 3G all the way. Now take the train. You'll have coverage (on Sprint and Verizon, at least) all the way to Orlando, but it's going to be 2.5G 1xRTT for most of the trip between West Palm Beach and Winter Haven. The only time you're going to see 3G is when the train is literally stopped at a station at one of the few medium-sized towns along the way. The moment you're a mile out of town (or sooner), it's 2.5G time again.
> That is becoming harder and harder every new model.
Not really. It's just Motorola phones that are crippled by design. With HTC and Samsung phones, at least, rooting is more like climbing over a low, wide wall that's lightly textured to make the experience a little bit unpleasant. Truth be told, I'll bet there are more than a few employees at HTC, Samsung, and probably Sprint & T-Mobile who'd LOVE to be running ads right now comparing Verizon and/or Motorola to Soviet Russia and East Berlin, but can't get management to sign off on them;-)
Fantasy Sprint/T-Mobile commercial:
Cute Google Android strapped face-down onto table that looks like a steampunk cross between a horizontal electric chair and a guillotine. Evil guy wearing military-looking uniform (with stylized 'V' logo) pulls out DroidX and cackles (screen wallpaper depicts Berlin Wall), grabs a thick cable with mean-looking plug on the end (like the ones used in the US for 3-phase 480v AC) and says, 'Vee have vays of dealing mit rootuzerz...' while plugging the cable into the Android's ass. Cut to hand grabbing Frankenstein-style knife switch, engaging the power, and a buzzing, high-voltage type noise that just happens to resonate in a way that sounds like the word "Droid!" at the end of a Verizon commercial being yelped in pain.
> Guess that'll teach ya to buy GSM only and direct from the manufacturer.
And have no 3G data service in a shockingly large part of America that isn't even particularly rural (the parts where you might have 3G service if you were to go stand on the roof of your house and orient the phone *exactly* the right way, but can forget about indoor service -- even next to a window. It's a particularly feast-or-famine problem with T-mobile. Due to their spectrum issues, there are quite a few places where the next step down from HSDPA/HSPA+ is GPRS (no EDGE).
For the most part, if you have Sprint or Verizon, you're going to get at least ISDN-speed 1xRTT data just about anywhere in the country that's within a mile of the nearest paved road, and have decently reliable 3G EVDO service just about everywhere you're likely to care about unless you're a park ranger.
> And how, exactly, does failing to sell a significant number of phones drive ANYTHING forward?
By making sure that any new phone with a processor slower than a Gigahertz, a display smaller than 800x480, or Android older than 2.1, would be laughed at and fail spectacularly. It was a way to forcibly make the handset makers & carriers skip over what would have otherwise been 2 or 3 short-lived incrementally-upgraded generations of phones and go straight to Nexus One specs.
Google did an even bigger favor. By making it utterly futile for carriers like Sprint & Verizon to even *contemplate* trying to EOL their first round of phones and roll out a second generation with incrementally-faster processors after Christmas (they would have been laughed at), it practically forced them to pay more attention to their 2.1 upgrades in a desperate effort to keep their aging phones (still being sold new, at massively discounted prices) halfway viable. Left to their traditional practices, they would have just slapped 1.6 onto the Hero & Droid Eris, then EOL'ed them after Christmas (or sooner), and rolled out a second generation of slightly warmed-over successors with more or less the same hardware inside, but running at 700MHz instead of 528MHz, and possibly with a little more memory... shipped with 2.0, and 2.1 promised for "summer".
In short, the Nexus One raised the bar any Android phone had to meet in order to be taken seriously, and in that respect it succeeded wildly. All we need now are whispered rumors of a "Nexus Four" (think: "two squared -- 2 cores x 2GHz = 4000mAH battery") to force the arms race to its next logical level.:-)
Slashdot Mirror
> Hint: look at Amtrak. It costs significantly more to travel per mile on Amtrak than it does to fly;
> it's slower, more cramped, less friendly, and you've got to wait longer.
Yes, it's more expensive to travel on Amtrak than to fly (at least, if you get a room), and it's a lot slower. However, there's no way in HELL it's more cramped than flying. Even in a roomette shared with one other person, upper bunk stowed, you have more real room than you'll ever have in first class on any domestic airline. Ditto, for seats in coach.
As for waiting... well, yeah... you might have to wait longer. But the quality of waiting for a train is different from waiting for a plane. When the train shows up, you can walk aboard without even having to shut down your laptop, or "discontinue use of approved personal electronic devices." If you're tethered to your cell phone, you might not even lose your current IP address unless you voluntarily disconnect at some point.
Amtrak is far from perfect. In fact, it's a pretty sad excuse for a passenger rail network. But even in its imperfect, sad state, it's nicer than the hellish ordeal flying has become over the past 10 years. Given a choice between a 3 hour trip in a shiny new overcrowded city bus, or a 5 hour trip in a slightly-tattered, but roomy and comfortable limo, I'll take the slightly ratty limo.
> Tampa and Orlando, a grand distance of 85 miles, or about 90 minutes driving.
On its own, you're absolutely right. A high-speed train that goes nowhere besides downtown Tampa to Orlando International Airport (with Disney sort of a bus ride detour along the way) is stupid. Where FDOT has a chance to redeem itself is if, by some miracle, it decides to build tracks suitable for "true HSR", but buys TRAINS capable of running at 125+mph between Tampa and Orlando, and equally capable of leaving those shiny new tracks near Auburndale and heading south to West Palm Beach, Fort Lauderdale, and Miami along the CSX tracks that already exist (possibly double-tracked, and improved a bit to allow 110mph operation for most of the way between Sebring and West Palm Beach, and 70-80mph the remainder of the way down to Miami). THEN, they make sense.
For the past 20+ years, FDOT's own studies have basically come to two conclusions: HSR will always hemorrhage money, and 90-110mph intermediate-speed rail will make money hand over fist. Why? The market is already here now for fast rail travel between South and Central Florida. Driving from Miami to Orlando sucks, and driving to Tampa sucks even more... but not *quite* badly enough to endure the hell and misery of flying. At ISR speeds, the corridor between Auburndale and Miami already exists; it's even double-tracked to passenger-rail standards from WPB to Miami. So, why would HSR hemorrhage cash? Because it would cost about $5-10 billion to build a really, really nice ISR passenger rail network connecting Miami, Orlando and Tampa... and $50-100 billion to build True HSR(TM) every last inch of the way down to Miami. $5-10 billion can be rationalized... it's basically $250-500 per Floridian. If nothing else, it would spur cross-state tourism like nothing the state has ever done.
And yeah, people would step off the train and grab a rental car, just like if they flew. The difference is, instead of spending 3 hours running and "actively waiting" (being forced to do nothing besides sit or stand and wait for permission to do something), they'd spend the same amount of time eating, drinking, or otherwise enjoying the first or last part of their trip. Ideally, passengers could do the rental car formalities on the train, and walk off at their destination with their rental car's keys in hand.
As for the HSR part between Tampa and Orlando, in an indirect way, it DOES still sort of make sense, even if the rest of the line is ISR. Why? Well, the tracks from Miami to WPB are already owned by FDOT and used by Tri-Rail (commuter rail). The tracks from WPB to Auburndale (halfway between Tampa & Orlando, about 5 miles south of I-4) are owned by CSX, but almost useless for freight. The tracks between Auburndale and Jacksonville all but been sold outright to FDOT (in fact, a segment has). HOWEVER... the tracks between Auburndale and Tampa are very, VERY valuable to CSX... and CSX has made it clear they won't give them up, or even share them, without a fight. And in Florida, they can fight dirty. It's right in the constitution... railroads can condemn adjacent land to their existing right of ways for basically any purpose without limit. So, if FDOT tried to use eminent domain to take the northern 50 feet of CSX's corridor, CSX will use eminent domain to take the next 50 feet or more to the south. And a few cities, including Lakeland, would be very VERY angry if FDOT induced CSX to do that. So FDOT doesn't dare.
*THAT* is why it sort of makes sense (at least, purely from the perspective of Florida's self-interest) to build the new tracks down I-4. Now, before the Feds showed up with a dumptruck full of money, FDOT's plans were a bit less ambitious... basically, a new track down I-4 to somewhere around Auburndale, then CSX the rest of the way to Orlando. But since the Feds are now eating a huge chunk of the cost, it's cheaper for Florida to just humor the feds and build 100% HSR all the way to Orlando.
NOW, IMHO, FDOT has the Orlando end all wrong. Instead of building
> and it's free- something that ought to appeal to poor starving college students.
Don't forget, if you're a college student, Microsoft software *is* practically free. Also, the "poor starving student" meme is grossly exaggerated, at least in the US, if you basically count full-time students at a major university (private OR public) who live on campus. Even if they ARE legally penniless, their cash-flow problems aren't going to kick in until after they've graduated and have to start paying back their student loans and maxed-out credit cards.
> but I do not think Linux is such a terrible operating system that it would see no use whatsoever
I think their statistics are skewed by only counting "Linux" if that's the one and only operating system on the user's main computer. Round up some students majoring in computer engineering or computer science, and I think you'd find that close to 100% of them have a small pile of laptops and/or additional partitions with Linux on them. They just aren't running Linux as their One True Daily Operating System.
> No ... no one bothered counting those 8 people because they are statistically irrelevant, stop bringing them up like they matter.
Er, I think there are quite a bit more than 8... http://forum.xda-developers.com/forumdisplay.php?f=735
> It was not patented. As such, the manufacturing will not go to China, but to Japan (who will then take it to China).
> Personally, Corning has earned my disdain. At this time, I will quit buying from Corning.
Er... why? If anything, keeping it a trade secret would have been a calculated gamble based on the apparent lack of real-world applications for it 60 years ago and lack of anticipated applications for it within the next 20 years.
Besides, even without patents and trade secret law, there's trademark law. At the end of the day, there will inevitably be factories in China making similar glass... but we all know that none of them will ever duplicate it exactly, even if they could, because the temptation will always be there to cut corners in some way people are unlikely to notice until it's too late. All Corning has to do is wait for someone to use the name of Gorrilla Glass in vain, and unleash the lawyers on them.
Could companies in Europe and Japan duplicate it? Sure. But they won't, because their manufacturing costs are as high (or higher) than Corning's, and they don't have the benefit of Corning's additional research on how to manufacture it cost-effectively. By the time they had it ready to sell, they wouldn't be particularly cost-competitive anyway. Like it or not, Corning pretty much owns the specific market segment for "expensive, exceptionally high-quality specialty glass for niche uses".
> Those places, there is actual broadcast TV that people can easily watch (without being dependent on the data bandwidth.)
Part of the problem is that American HDTV uses 8VSB instead of COFDM for data transmission. Oversimplifying a bit, 8VSB is great if you want the longest range over flat terrain possible AND assume your audience has properly-grounded directional antennas... but it's difficult to receive in a moving vehicle (even WITH a good antenna), and nearly impossible to receive by a pocket receiver with stub or internal antenna. In contrast, COFDM has a lot less "fringe" range, but has fairly minimal antenna requirements if you're within a few miles of the tower.
The net result is that in Europe and Japan, a pocket-sized HDTV tuner is likely to work well in urban areas. In America, you'd be lucky to get a viewable HDTV signal anywhere besides maybe Weehawken (New Jersey... high on a cliff, line of sight to the antennas on the Empire State Building ~3 miles away).
In other words, it was one of those, "Oops" moments when a new standard got declared by law, and nobody realized the oversight until it was too late to do anything about it. Or so the TV industry claimed, right before rendering every HDTV sold up to that point effectively obsolete for its intended purpose by virtue of not supporting HDCP...
> It's most telling that Palm is flatlining and Windows Mobile has lost half of its already
> meager market share in the past year.
Are you counting people who own phones that were sold with Windows Mobile, but are now running Android (like the HTC Touch HD2)? The HD2 debacle will go down in tech history as one of Microsoft's worst marketing/business decisions in history. Here's a phone that was eagerly embraced by Microsoft's few remaining enthusiasts, even as their friends and peers ran for the door marked "Android", only to get its owners metaphorically kicked in the balls by Microsoft on what was probably the lamest pretense for non-compatibility *ever* (it had four buttons instead of three).
Microsoft could hardly have done a better job of driving its few remaining friends into the Android camp if they'd personally rebranded MSDN as an Android portal & given a free Nexus One to everybody who attended a Microsoft event in 2010.
My apologies. I'd read a few summaries earlier in the day, and was under the impression that it was up to the app itself to persistently keep track of the app's licensing state if it didn't want to check every single time.
Now, if only Google could either bully HTC, Samsung, Motorola, and the rest into leaving HID enabled when they build the kernel (so AOSP could implement the remainder at some later point without needing a new kernel), or come up with some good way to allow users with 100% compiled-from-the-ground-up (kernel and all) AOSP builds to run protected Market apps. Say, by distributing the DRM support as a loadable kernel module augmented by the DRM built (by definition, since that's what officially differentiates "SD" from "MMC") into every (micro)SD card ever sold.
There were (at least) two fundamental flaws with the original Android Market protection scheme, neither of which appears to have been rectified by this change (besides possibly to make matters worse for end users):
* As everyone has already noted, lots of people around the world with Android phones can't actually buy apps from Android Market, EVEN IF they have a Mastercard/Visa/AMEX card with dollar-denominated account. That's just plain fucked.
* You can't officially purchase and run protected Market apps if your phone is running an unblessed "Developer" kernel. Of course, there's not a single goddamn phone from HTC, Samsung, or Motorola with Google-blessed kernel that has BlueZ Bluetooth HID profile compiled into it, so it's impossible to build your own kernel with it enabled without being formally exiled from 99% of commercial Android apps. At least, unless you crack them. Any DRM scheme that forces legitimate users to crack apps they purchased in order to use them is fundamentally broken, especially when there are still gaping holes in Android phones that need a customer kernel to fix.
As for "developer's option" whether or not to cache, let's be honest... at least half the developers publishing commercial apps don't have the slightest clue in HELL how to implement a secure caching scheme, and they aren't going to purchase a proprietary one that demands more money up front than they're likely to earn from the app's sale. So, anybody care to guess what's going to happen? Most apps in Market are going to end up checking the server every goddamn time, because the alternatives are too hard/expensive for most Android publishers to deal with. IMHO, Google got THAT part EGREGIOUSLY wrong. They should have distributed the Android DRM module themselves, and made it free & easy for publishers to do cached checking, but left it difficult and minimally-documented how to bypass that caching and check the server every time.
I love Android. I really do. But it's so incredibly frustrating when Google turns around and fucks things up in ways that CAN'T be fixed by end users with access to Android's sourcecode... usually, mistakes that are almost incomprehensible given the amount of in-house talent and expertise Google has available to it. At times, Google actually manages to make even *Microsoft* look coherent and customer-focused.
> Also didn't we have all those things about 100 years ago?
Exactly. If anything, it could almost be argued that the pollution in late 19th-century Britain, France, and Germany (and parts of America, for that matter) were noxious/toxic enough to make the most badly-polluted square mile of China look like the Garden of Eden by comparison. At least people in China don't have to rely on wood and coal-burning stoves & fireplaces for cooking and heating ON TOP of the pollution being produced by factories (at least, urban factory workers who live amidst the worst pollution) don't.
As a species, humans are easy to kill individually, but surprisingly difficult to effectively exterminate. The dinosaurs didn't have preserved food, hydroponics, artificial lighting, and global distribution networks, so when the skies went dark and 99% of photosynthesis shut down for a few years after the impact event, they were screwed. A similar event would be an unprecedented human tragedy, but the likelihood of enough humans surviving to repopulate the Earth eventually is practically assured.
Someone please mod this up. The games in danger of fading into oblivion aren't Pac Man, Dig Dug, and Centipede... or even the more obscure games from that era, or even the 90s. The games in danger are the DRM'ed MMORGs that have no existence independently of the company hosting them online.
Historically, most preservation is done by people who aren't the official owners. Just look at Hollywood -- execs used old, only-existing-copies of silent films from the 20s to start bonfires at beach parties for years. Most of the old films we have copies of came from private collectors, many of whose copies were technically (if not outright) illegal for them to own.
It's one reason why librarians are more than slightly disturbed about predictions that ebooks will replace bound, printed volumes. It's not that they're anti-technology... it's the fact that books have a way of persisting and remaining accessible long after whomever owns the copyright has lost interest in them for any purpose besides random windfalls due to infringement lawsuits. In contrast, DRM-protected content can go away for reasons ranging from active neglect all the way to intentional retraction. It's pretty much a given that cracking 1024-bit encryption will be do-able 50-100 years from now, but the danger with digital content is the combination of its short media life and persistence of copy protection. If ebook readers protect content not only by encryption, but by copy-protection as well, and the media doesn't retain its integrity long enough for the DRM itself to be crackable... well, the problem is obvious.
I think it's not so much a matter of not knowing about this as a potential vulnerability, as it is a case of the hardware necessary to pull it off suddenly becoming cheap and affordable to just about anyone with the slightest interest in doing it.
Perfect illustration of "exploit" that becomes possible due mainly to falling prices:
My best friend owns two Blu-Ray players. One is Region "A". He bought it for $99 the day after Thanksgiving last year. The other is Region "B". He paid around $160 for it, including shipping, from the guy in Hungary or Romania who was selling them on eBay. Both are perfectly happy to feed 1080p24 video via hdcp-protected hdmi to his TV. If he cared, I'm sure he could buy a Chinese Region "C" Blu-Ray player for another hundred bucks or so, plus maybe another $50 for a HDMI switcher. Region coding only works as long as players are too expensive to just go out and buy one from every region you care about.
It depends. From what I remember (teen in the late 80s), at least in the Apple //e, Atari 400/800, Vic20, TI99/4A, C64, Atari ST, and Amiga era, most games didn't start out as a brilliant high-level concept. Someone discovered a cool video hack, and managed to build a game around it. That was part of the reason why ported games on any platform almost universally sucked... whatever it was that made them look GOOD on their native platform didn't exist (or was moot and no big deal on the new platform), and forced to stand on their own, they fell flat on their face.
The games that lasted were the ones that did, in fact, transcend the video hack that made them possible in their first implementation. Jumpman is still fun, warts and all. In contrast, most of us wouldn't waste the erase cycle on a flash drive for a pirated copy of Incredible Mission, because the only thing that ever made it interesting was the relatively high-resolution (for the time) graphics and 4 seconds of digitized speech at the start. A hundred years from now, at least a few people will still know what Zork was. A few might even have played it in some context. Nobody will care about playing Elvira, Mistress of the Dark on an Amiga emulator.
Actually, you wouldn't even need a vector graphics monitor. Given a decent 20" glossy 1920x1080 LCD (possibly in portrait mode) capable of 120hz native updates with subpixel control via hdmi or dvi, combined with triple-buffered video emulating the bloom, defocusing, and (in the case of a color vector display like the ones used for Defender and Tempest) fringe artifacts caused by misaligned shadow masks, you could emulate a vector display with more or less perfect accuracy.
Emulated games looked like crap on 1024x768 LCD panels because they didn't scale well, and didn't offer enough raw control over the rendered video to recreate the appearance of an interlaced CRT. When you're talking about a display that can do 4x oversampling with respect to both resolution and framerate, and a computer fast enough to emulate the phosphor behavior of a CRT, you can achieve nearly perfect emulation. Hell, if you made a Vectrex-sized LCD with the pixel density of a Retina display, you could probably emulate a monochrome Vectrex down to the exact bluish-white raster on a grayish background. We're not *quite* at *that* point yet, but in another 5-10 years, we absolutely will be. At that point, having physical vector displays ceases to matter for the experience of reliving the past, and really matter only to act as a reality-check against creeping "improvements" beyond that point to the emulated video algorithm that make it look better, but destroy its historical accuracy. Does anybody *really* want to emulate the appearance of burnt-in phosphors, at least while actually *playing* an old game?
Let's face it... most 3-4 year old games at mall video arcades looked like shit thanks to years of burn-in, bloom, and slowly-dying electrolytic capacitors on the circuit board (not dying as quickly as "turn of the century" bad caps, but still visibly degraded compared to when they were new).
> Because the actual GLP'd kernel code is available, just without the proprietary drivers.
> There is no violation because the source code is available.
Actually, no. It *is* pretty much a clear-cut open and shut GPL2 violation. FSF can argue with Linus over whether or not loadable kernel modules are or are not part of the kernel proper (and thus subject to requirements that their source be released), but I don't think there's *anyone* who's going to argue that what HTC does is OK.
The problem is, you can't just go and sue someone for "violating the GPL(2)". You have to prove in court that:
1) You have standing to bring the case (ie, you're one of the people who collectively own the Linux kernel's copyright)
2) The court you've chosen is the proper venue to pursue the case.
3) The code you contributed is in the kernel they shipped.
4) The source files they released were legally inadequate to fulfill their obligations as a licensee under the GPL2
5) You suffered real harm due to their actions.
Getting past step 1 could easily cost tens of thousands of dollars and involve multiple court appearances. Rest assured, the defendant's law firm is going to do everything they can to cast doubt upon your standing. If that fails, they're going to do everything they can to challenge your choice of venue (ie, the authority of the court to hear your case, and its appropriateness).
3? The easy part. Don't smile yet, because 4's going to be a bitch.
4) Have fun proving they violated the GPL2. Common sense might dictate it, but there's surprisingly little case law to actually cite one way or another because most lawsuits involving the GPL end up getting settled at the last minute & vanish from the legal radar.
5) This is the toughest of all. To get the grand prize you really want -- equitable relief granting a plea to force them to "go forth and sin no more", you have to prove that their actions have harmed you. The best-case here is probably if you own the device they shipped without the source and was unable to build your own copy of the kernel you helped develop and partially own because of their infringement. Of course, if you had to root your phone to get it into a state where it's physically possible to make use of such a kernel, you can bet they're going to throw every legal theory they can at you in the hope something will stick and get you classified as having "unclean hands" (ie, you're at least partly responsible for your plight). They might prevail, they might not, but they'll fight hard & fight dirty.
However, it doesn't end there. Suppose the judge agrees that they were totally wrong, harmed you & the larger community, and agrees to issue a court order demanding that HTC release the source to everything included in their kernel. You can bet that before anyone at HTC fires up a text editor to go to work on preparing the source, Qualcomm and everyone else who furnished those proprietary binaries (or info under NDA necessary to implement them) will have injunctions of their own to stop HTC from releasing source they don't have the right to release. If you're lucky, you might be able to force HTC to do what they should have done in the first place: re-implement them as proper loadable kernel modules, rebuild the entire kernel so it works with them, and release THAT... and do the same for everything going forward. Realistically, the likelihood of this happening fewer than 5 years from the moment you walked into the law firm's office to kick off the case is depressingly slim.
Oh... also... if the trial DOES drag on for years... don't let HTC's private investigator catch you using a different phone. Courts won't hear cases that are moot. If your claim that they've caused real harm to you rests upon being unable to build a kernel of your own for your phone (based on their infringing release), and they can demonstrate you haven't touched that phone in 3 years... well... let's just say it wouldn't be good. It wouldn't necessar
OK, here it is... the official flag of the Evil Empire:
http://img84.imageshack.us/img84/5555/evilempire3.jpg
Join the 'Droid Liberation Army. Fight the Power. Er, and stuff like that... ;-)
(Image released under the Creative Commons License)
I can't speak for the Droid, but the specific problem lots of people have building kernels for HTC phones is the fact that HTC compiles their proprietary binary drivers straight into the kernel itself, then distributes the whole thing as a big monolithic blob. When it's time to distribute the source, they go in, rip out the source to anything proprietary, and dump the rest into a tarball for download. The problem is, you need the functionality provided by those proprietary binary drivers to do use things like wi-fi and the camera, but because they're effectively raw, undocumented binary blobs, you can't easily build a new kernel that incorporates them into it. Worse, all of the furnished build scripts released by HTC fail, because they attempt to reference 'include' files that aren't... well... included.
The PROPER, Linus-blessed way to incorporate proprietary binary drivers into a Linux kernel is to distribute them as loadable kernel modules. By defining a clear public interface and drawing a clean line between what's proprietary and what's not, you can then create a buildfile that compiles the part you DO have the source to in a way that works seamlessly with the binary kernel modules to which you don't, and everything "Just Works". FSF can bitch about not having the source to the WLAN drivers, but if the choice is binary WLAN drivers or no WLAN at all, .ko loadable kernel modules are a pragmatic compromise that lets you keep the functionality provided by them without breaking your ability to rebuild the rest of the kernel yourself.
> You'd think the public would want to dash towards free calls.
Because, in America at least, there's no such thing as free/low-cost wireless service that's pervasive AND works with anything vaguely resembling a high-end Android phone. American wireless service is kind of like a nightclub with a high cover charge, but free drinks and all you can eat buffet once you're past the doorman & cashier. If you're paying Sprint $90/month for "unlimited everything", there's no incentive to screw with Skype unless you have to make international calls. And with most calling cards seemingly hovering at around 2.9c/minute for international calls, even those aren't worth the hassle and latency unless you're wearing a headset and watching synchronized DVDs together with someone on the other side of the world for hours at a time.
> You should totally spend some quality time with Blender and animate that. That could go viral in a heartbeat.
LOL, one of my undergrad majors was advertising, creative track. Copywriting, though... not art direction. :)
I *would* appreciate being credited as "Miamicanes" if someone with production skills chooses to make this into a real video, though :)
> You will never get modded up as you deserve. I think this has been the clearest description of the hypocritical
> thinking Android fans (notice didn't say fanbois...awful term) have vs. iPhone.
The difference is, we openly and routinely excoriate HTC, Motorola, and the others, bitch about their GPL violations to anyone who'll listen, and compare them to Satan's lovechild for even the smallest transgression. iPhone fans rationalize and justify Apple's behavior, and act like it's somehow shameful to demand full control of your phone.
Android owners bitch about the difficulty of building an Android distro from scratch without the cooperation of the phone's maker, usuall caused by things like HTC shitting monolithic binary kernel blobs on the curb, sniffing them a few times, and walking away satisfied instead of building their proprietary binary kernel drivers as loadable kernel modules the way they're REQUIRED to under the GPL (so new kernels can be built around them without losing the functionality provided by the .ko modules themselves).
It's not hard to imagine a few thousand angry Android users staging a protest in the Googleplex parking lot over some perceived betrayal of Android's open ideals. Try to imagine even a few dozen iPhone owners picketing on the sidewalk in front of Apple if AT&T somehow managed to push out an update that revoked root and forcibly reflashed a million jailbroken iPhones. It's almost inconceivable. Even if there were a few dozen angry iPhone owners, they'd be drowned out by the ocean of Normal Users(tm) bleating about how they shouldn't have jailbroken their iPhones in the first place, because jailbroken iPhones makes Steve Jobs sad.
> Hell, on a drive from Houston to Dallas or Austin I can tether the thing to my laptop and stream Grooveshark
> with barely missing a beat, and that is a bit of a rural area between here and there.
T-Mobile, like Sprint, Verizon, AT&T, and everyone else, deploys coverage in rural areas first to highway corridors between their "real" coverage areas. It's a lot cheaper to throw a tower every couple of miles next to a major interstate along the "Texas Tee" than it is to keep building those towers further away from the road. I can drive from Fort Lauderdale to Naples along Alligator Alley and enjoy EVDO the entire way... but if I pulled off the road at one of the rest stops, got into a canoe, and went two or three miles north or south, I'd be lucky to have voice service that worked.
You can see this pretty acutely if you take Amtrak from South Florida to Orlando, and compare coverage along the way to the coverage you have while driving along the Turnpike. On the Turnpike, you'll have 3G coverage from pretty much everyone -- Sprint, Verizon, T-Mobile, and AT&T -- all the way to Orlando. It might drop to 2.5G for a few miles ~50 miles southeast of I-4, but for the most part it's 3G all the way. Now take the train. You'll have coverage (on Sprint and Verizon, at least) all the way to Orlando, but it's going to be 2.5G 1xRTT for most of the trip between West Palm Beach and Winter Haven. The only time you're going to see 3G is when the train is literally stopped at a station at one of the few medium-sized towns along the way. The moment you're a mile out of town (or sooner), it's 2.5G time again.
> That is becoming harder and harder every new model.
Not really. It's just Motorola phones that are crippled by design. With HTC and Samsung phones, at least, rooting is more like climbing over a low, wide wall that's lightly textured to make the experience a little bit unpleasant. Truth be told, I'll bet there are more than a few employees at HTC, Samsung, and probably Sprint & T-Mobile who'd LOVE to be running ads right now comparing Verizon and/or Motorola to Soviet Russia and East Berlin, but can't get management to sign off on them ;-)
Fantasy Sprint/T-Mobile commercial:
Cute Google Android strapped face-down onto table that looks like a steampunk cross between a horizontal electric chair and a guillotine. Evil guy wearing military-looking uniform (with stylized 'V' logo) pulls out DroidX and cackles (screen wallpaper depicts Berlin Wall), grabs a thick cable with mean-looking plug on the end (like the ones used in the US for 3-phase 480v AC) and says, 'Vee have vays of dealing mit rootuzerz...' while plugging the cable into the Android's ass. Cut to hand grabbing Frankenstein-style knife switch, engaging the power, and a buzzing, high-voltage type noise that just happens to resonate in a way that sounds like the word "Droid!" at the end of a Verizon commercial being yelped in pain.
> Guess that'll teach ya to buy GSM only and direct from the manufacturer.
And have no 3G data service in a shockingly large part of America that isn't even particularly rural (the parts where you might have 3G service if you were to go stand on the roof of your house and orient the phone *exactly* the right way, but can forget about indoor service -- even next to a window. It's a particularly feast-or-famine problem with T-mobile. Due to their spectrum issues, there are quite a few places where the next step down from HSDPA/HSPA+ is GPRS (no EDGE).
For the most part, if you have Sprint or Verizon, you're going to get at least ISDN-speed 1xRTT data just about anywhere in the country that's within a mile of the nearest paved road, and have decently reliable 3G EVDO service just about everywhere you're likely to care about unless you're a park ranger.
NASCAR!!!!! Argh!
> And how, exactly, does failing to sell a significant number of phones drive ANYTHING forward?
By making sure that any new phone with a processor slower than a Gigahertz, a display smaller than 800x480, or Android older than 2.1, would be laughed at and fail spectacularly. It was a way to forcibly make the handset makers & carriers skip over what would have otherwise been 2 or 3 short-lived incrementally-upgraded generations of phones and go straight to Nexus One specs.
Google did an even bigger favor. By making it utterly futile for carriers like Sprint & Verizon to even *contemplate* trying to EOL their first round of phones and roll out a second generation with incrementally-faster processors after Christmas (they would have been laughed at), it practically forced them to pay more attention to their 2.1 upgrades in a desperate effort to keep their aging phones (still being sold new, at massively discounted prices) halfway viable. Left to their traditional practices, they would have just slapped 1.6 onto the Hero & Droid Eris, then EOL'ed them after Christmas (or sooner), and rolled out a second generation of slightly warmed-over successors with more or less the same hardware inside, but running at 700MHz instead of 528MHz, and possibly with a little more memory... shipped with 2.0, and 2.1 promised for "summer".
In short, the Nexus One raised the bar any Android phone had to meet in order to be taken seriously, and in that respect it succeeded wildly. All we need now are whispered rumors of a "Nexus Four" (think: "two squared -- 2 cores x 2GHz = 4000mAH battery") to force the arms race to its next logical level. :-)