The root of Sprint's thrashing problem is the RIL programming that prevents simultaneous wimax and CDMA (EVDO or 1xRTT) data sessions. Now, when a Sprint phone attempts to connect to Wimax, it first terminates any live data session (EVDO or 1xRTT) before it even tries to authenticate to Clear. Likewise, if the wimax connection fails (or is failing), the phone will first fully terminate the wimax session before it even tries to establish a CDMA data connection.
What Sprint really needs to do is give us an option to keep a hot EVDO or 1xRTT data session as a live backup whenever wimax is active, and have the phone establish a fairly conventional multilink VPN with dynamic failover to a VPN server somewhere (favoring wimax, but transparently rerouting data to EVDO or 1xRTT the moment a failure occurred). That way, applications and the outside world would both see a stable IP address, regardless of how ugly things got between the phone and towers.
Technically, this is entirely possible to do with the phones we have RIGHT NOW. In all of Sprint's current phones, the wimax and wi-fi radios are on the same daughtercard (in fact, I think they're the same chip). If it weren't possible to do simultaneous 1xRTT and wimax, it wouldn't be possible to do simultaneous wi-fi and voice calls, either (voice calls and 1xRTT work the same way). It's a shame the radios aren't as hackable by end users as Android itself is. Otherwise, I'd give it 3 months, max, before custom modems started to appear that could use thirdparty VPN services together with simultaneous CDMA + wimax data connections to achieve instant failover. Battery life would probably suck, but at least you'd never find yourself swearing at Connectbot because you forgot to disable 4G before connecting, then at some point the phone spontaneously decided to kill your SSH session so it could try and connect to a new wimax tower it just noticed....
^^^ What he said. When Qualcomm introduced CDMA in the early 90s, it was about as close as you could get to black magic. Engineers from Nokia or Ericsson went on record officially calling it a fraud, and ultimately had to eat their words a couple of years later when they were proven wrong.
The coolest thing about CDMA is that you can literally do things like throw down a low-power cell in the middle of a congested area (like a mall, university, stadium, etc), add it to the network, and watch the congestion just kind of automatically sort itself out and self-optimize the power levels. CDMA networks can grow organically, and in an adhoc manner that would be inconceivable with legacy GSM (not UMTS, because UMTS is CDMA)
>Or if you like to actually own a phone and can swap out SIM cards to change phone numbers and carriers quickly.
CDMA has a perfectly good standard called R-UIM that's a superset of GSM SIM, and an optional subset of USIM, that serves the exact same purpose. Unfortunately, Sprint and Verizon never bothered to support it. Elsewhere in the world, CDMA phones and networks are as mutually interoperable and carrier-agnostic as GSM phones.
Once you've hit the home button and launched the camera app by finding its icon and clicking it. As opposed to the barbaric, uncivilized way Android owners do it... by pressing a (gasp) dedicated camera button;-)
I don't remember the exact year offhand, I guess it might have been closer to 15 years ago. I just remember that the car was manufactured and bought the final year CFC-using air conditioners were legal, and the repair happened about 5 years later. In retrospect, I probably should have taken the car to get fixed by some random mechanic in his driveway in Hialeah (who could have refilled it with used refrigerant recovered from a junkyard car), but I made the mistake of taking the car to the dealer and got completely raped, just like a few million other Floridians who had no idea at the time how expensive something that was historically a minor band-aid repair job you did to older cars in their final year or so of life before trading them in had just become (the AC in question wasn't completely broken... it had just lost enough refrigerant over the years to eventually start icing up regularly, and really needed only a fairly small amount to top it off).
So, how exactly did a level of CFCs that are now a fraction of a percentage point of any modern-era level -- to the point where they're now banning fucking *asthma inhalers* -- somehow cause the largest "ozone hole" in recorded history? This seems to suggest that we've either gone completely overboard trying to banish every last trace of something useful in an act of complete futility and political masturbation, last year was an epic flaming catastrophe with millions of halon-extinguished fires (since that's pretty much the only significant source of CFCs left), or CFCs don't have nearly as much to do with it as originally speculated, and the "holes" are going to keep occurring regardless of whether or not CFCs continue to exist.
This is not to deny that the phenomenon exists, or that CFCs have no impact, but rather to point out that the worldwide banning of CFCs might have ultimately caused a phenomenon that's going to occur anyway to now be 99% as large as it might otherwise have been instead of 100%, and call into question whether that small improvement was actually worth the expense and misery of its opportunity cost. As someone who got hit with a $1,800 AC repair bill ~9 years ago for something that should have been a $65 refill (because they had to basically tear out and replace the car's entire air conditioner), I have a bit of an axe to grind over this matter. Frankly, I'm not convinced that I've gotten $1,735 worth of added ozone value in return for my investment.
There's a difference between arguing that CFCs shouldn't have been phased out at all, vs arguing that the extreme and expensive measures taken to banish them almost overnight regardless of cost or benefit instead of simply requiring that new equipment be CFC-free going forward were worth it.
> People seem to want symbolic icons that represent the programs they want to run; > they don't want to look through a long menu and read a bunch of text.
Oh god, no. Please don't remind me about Lotus SmartSuite's Hieroglyphics from the mid-late 90s.
Give me a nice, recognizable icon AND text, so I can recognize the icons I care about frequently, then find the remaining functions without having to play "guess what this is supposed to be symbolic of".
It's still stupid for them to not support 1700MHz uplinks. Even if AT&T *does* buy T-Mobile, it's not like they could instantly absorb all of their customers into AT&T's existing 3G network overnight without making life suck even MORE for existing AT&T customers. If the iPhone 4s phones sold by AT&T supported 1700/2100 (AWS) UMTS, and AT&T closed the deal, they could instantly add T-Mobile's existing 1700/2100 HSPA+ sites to AT&T overnight with little more than a PRL update.
That said, is the iPhone 4s physically INCAPABLE of 1700MHz uplinks, or is it merely not CERTIFIED for them? I find it hard to believe that Apple has somehow managed to buy millions of WCDMA baseband processor chips that are LITERALLY silicon-incapable of 1700MHz operation. Personally, I suspect the iPhone 4s's bare-metal hardware is perfectly capable of 1700MHz AWS, but won't be submitted to the FCC for certification as such unless the AT&T merger falls through.
The main risk to sitting upright is aortic dissection. Apparently, when the body comes to a rapid stop in a forward-facing upright seated position, your internal organs can shift forward before coming to a stop and cause the aorta to tear enough to cause you to die from uncontrollable internal bleeding -- possibly minutes or hours AFTER you've gotten off the plane and are busy celebrating your survival.
Not quite right. South Korea USED to be 100% CDMA, but a couple of years ago the carriers decided to skip EVDO and go with W-CDMA/UMTS instead. The result is that South Korea's cellular network is like Canada's and semi-rural Australia's -- CDMA2000 for legacy voice and 1xRTT data is available everywhere, 3G is strictly UMTS, and some new phones sold there can't do CDMA2000 at all.
Or, if you're a Sprint customer due for an upgrade, buy a Motorola Photon. It does both GSM and CDMA. Just make sure you get the SIM unlock code from Sprint before you leave, and be aware that the SIM is used ONLY for GSM identity (in theory, Sprint could have used it with a USIM card capable of operating as a CDMA R-UIM... but they didn't), so you won't be able to casually use it on any CDMA network in China (except as an expensively-roaming Sprint phone).
If you're trying to make a blanket, universal declaration, yes it's evil. In this specific context, though, you have the database of a worldwide book retailer bought by another one that's more or less equivalent, who plans to use it for more or less the exact same purpose in the same way as Borders.
Now, things might be different if the mailing list were purchased by some religious group who planned to scour it for "unwholesome" purchases to identify "sinners" and make their lives miserable... and that's why there should be a degree of judicial oversight and a legal requirement that any such transfer be approved by the court, and only to a purchaser with the intent and means of using it in the same way for substantially the same purpose as before. Barnes & Noble clearly falls into that category. Amazon mostly does (no actual brick & mortar stores). Wal Mart or Buy.com would be a slippery case. The New Reformed Church of Jesus Christ the Vengeful Redeemer (and its 17 members, not including the psychotic preacher's family) from somewhere in rural Mississippi would be an obvious example of the other extreme. The hard part is drawing the line in a way that makes it easy for someone like B&N to buy it, and impossible for someone like the NRCJCVR to buy it.
A reasonable compromise might be a law that placed the data in escrow, and allowed its purchaser a single opportunity to contact the individuals via email and allow them to either formally opt in and allow their data to be handed over to the purchaser, or do nothing and have their data be permanently destroyed 6 months later. With a further rule that no single database can be sold to more than 3 potential purchasers, and that anyone receiving the first email can explicitly opt out of both the current offer and all remaining ones with a single click.
I can see why some are objecting to this specific transfer for the sake of having a consistent policy, but hysteria over it is just making the broader cause look silly.
AT&T was able to catch iPhone users who tethered because all but one of the iPhone tethering apps hijacked IOS' built-in tethering support, which was trivially easy for AT&T to track due to the way it worked. An app like EasyTether for Android is effectively untraceable by carriers if you use it to establish an IPSEC VPN, because even with deep packet inspection it's indistinguishable from an android app creating the same VPN for its own use. If necessary, we can raise the stakes and use SSH to resurrect SlipKnot to our own servers. Or masquerade it as an RDP back-channel audio stream.
There's a difference between carriers being able to cast a net and catch unsophisticated iPhone users who downloaded some point and click tethering app, and being able to catch people who are hellbent on sliding under the radar and have the technical skill to pull it off.
Pair-bonded IDSL is basically two IDSL lines aggregated as PPP Multilink. It won't do much for a single user using something like RDP or VNC, but for web browsing that involves one request for a page followed by a burst of requests for separate files, it effectively doubles it.
If you're willing to get down & dirty with Covad, pull out the checkbook, and let them work their magic on dry copper pairs leased from the ILEC, they can do even more impressive things and go places where mainstream mass-market ISPs are afraid to tread. I've read of people in rural Tennessee who've gotten 460kbit and more 25 miles from the nearest CO. The catch is, it's not cheap. They're basically building a private ISP for you alone by leasing rackspace at a central office, and pulling off some fairly expensive DSP magic (possibly with repeaters along the way to regenerate the signal) to get you the fastest data possible.
It's not mainstream, because most "rural folks" who aren't wealthy New Yorkers living in a cabin in the mountains because it's cute and cool can't afford to spend several hundred dollars per month for what's basically commercial-grade internet access, but you'd be amazed at what you can get if installation fees in the neighborhood of $1k-5k and monthly service fees between $250 and $500 are tolerable. I think the pair-bonded IDSL route is the most popular solution in Tennessee, because the way phone service is regulated there makes ISDN dirt cheap compared to the rest of the country, so it's not prohibitively expensive to just order more lines and aggregate them in parallel with PPP multilink the way people used to shotgun 56k modems in the late 90s.
> Or how would you recommend depositing a check/cheque to such a bank?
If you don't want to spend significant amounts of money on internet service? With a PC, a scanner, and a dialup modem. Worst-case, you can probably run the Android emulator in wifi mode (not sure whether the emulator can use the PC's webcam, I've never actually tried it).
Well, strictly speaking, they didn't "block" them, they just didn't allow them to be shown in Android Market. They made it non-easy for unsophisticated users, but didn't actually make it *hard* for regular users the way AT&T did.
Now, if they started poisoning DNS to make their domain appear to be invalid, or started to actually intercept and mangle http requests to their web site, that would be much more incrementally-evil and condemnation-worthy. On a scale of 1 to 10:
Filtering from Android Market: 2 DNS poisoning: 7 Http filtering: 8 Disabling installation of non-market apps: 9 Disabling ADB: 10 (I think this is actually forbidden by Google's licensing) AT&T in General: 17 AT&T's management: off the scale. Alderaan is fucked, and only the FTC can save the rest of us.
If Sprint DOES end up with exclusivity, I can almost guarantee it won't extend beyond December 1. Honestly, I wouldn't expect it to last beyond November 15th. Sprint will probably get to have it to themselves for four weeks while Foxconn builds more for AT&T and Verizon, then Apple throws open the floodgates to make as much money as it can before Christmas. There's no way in HELL Apple would risk losing Christmas sales to people who'd prefer an iPhone, but will take a high-end Android phone if it means having it to play with on Christmas Day.
For the sake of accuracy, the only carrier known to have ever done that in the US is AT&T, and they appear to have quit doing it for new phones going forward from the Infuse, and supposedly are unlocking older phones as they roll out periodic updates over the next few months. Now, whether AT&T will KEEP leaving them unlocked if it loses its fight to buy T-Mobile, and quits trying to publicly pretend that it's non-Evil, is anybody's guess.
> Now you understand why Samsung just hired Steve Kondik, founder of the Cyanogenmod project. > They need someone like him very badly.
You're absolutely right. Actually, Steve will help Samsung a lot, because for basically the cost of one happy full-time employee, they've effectively outsourced the long-term maintenance of their phones' firmware to dozens to hundreds of enthusiastic, highly-skilled unpaid volunteers (many of whom would be VERY expensive to hire for real as full-time employees). Samsung has NEVER been good about supporting phones with updates after they've shipped. Cyanogen went a long way towards fixing that, but had a problem -- there were hardware issues that just couldn't be easily solved without access to the proprietary bits of source that Samsung couldn't hand out to members of the public. Now that Steve's an employee under NDA, they can give him the keys to the castle and let him freely build flawless Cyanogen-optimized kernels for Samsung's phones, and leave everything else up to the community.
> Except perhaps people who live outside the coverage area of cable and DSL service, where nobody offers >"high-end internet access at home", where the best options for home Internet are satellite and fixed 3G/4G
Please, enlighten me. Name one Google'able address anywhere in America where there's solid 300kbps+ EVDO coverage from Sprint, but cable internet or pair-bonded IDSL of comparable speed or better is unavailable at any cost. No, 153kbps 1xRTT doesn't count, nor does the ability to mount a directional antenna on the roof and get fixed wireless service through Wimax. I'm talking about literally whipping out an Evo indoors and getting anything better than sub-ISDN speeds if you're lucky. People who practically live in the shadow of Sprint's towers can barely get 4G to work indoors, and you honestly think you're going to have viable 4G service on a PHONE somewhere that's so far out in BFE, you can't even get pair-bonded IDSL or cable internet?
iPhones and Android phones aren't for frugal people who want to spend $12/month to make emergency calls. They're for people who live online 24/7, and would dump satellite TV and their landline phone long before they even contemplated giving up unlimited wireless data and top-tier broadband.
Compared to what? Sprint has plenty of warts, but price (at least, for individual customers with no family plan and exactly one phone who'd burn through AT&T and Verizon's caps within a matter of days and are perfectly cool with $69.95 + $10 for 450 minutes of peak airtime, and more or less unlimited everything else) isn't really something I'd classify as one of them.
> Providing better 4G caps than the competition.
Root your phone like everyone else, and the 4g hotspot caps are meaningless. Sprint chose that specific group very carefully -- the nontechnical users who actually go out and PAY for an official hotspot plan are almost exactly the same group who'll try to use their phone as their one and only means of internet access.
Sprint isn't stupid... they know the overwhelming majority of users who root, reflash, and tether for free already have the most expensive cable or DSL internet access they can buy, and use it instead of their Sprint data service when they're at home just because it's faster and works better. To repeat: Sprint really doesn't care about users who tether once in a blue moon so they can get online with their laptop at an airport somewhere while waiting for a plane. Sprint passionately cares about users who try to use tethering as a substitute for real internet service and 21st-century dialup.
Nobody who has high-end internet access at home is going to screw around with torrenting from a tethered phone, because it would be slow, suck, and annoyingly cause most of your incoming calls to end up going straight to voicemail. Likewise, statistically nobody with the means to tether is going to stream lots of HD video, because it's not free -- users who tether for free rip their content from Blu-Ray, convert it to.mkv, copy it to their 32-gig Class 10 microSD cards, and watch it from THERE. Sprint is one of the few carriers who understands that the users who can most easily subvert any controls they try to impose are likely to be the ones who fall towards the lower end of total monthly data use, simply because those users have better ways of getting online anyway.
> Upgrading network capacity.
No arguments there. Sprint definitely has plenty of room to improve in that regard.
^^^ Just to clarify one point that I just thought of... TowerCo owns the land/lease and the tower itself. TowerCo does NOT own the spectrum licenses or operate the actual equipment. It's the equivalent of a colo facility where carriers (including Sprint) rent the equivalent of neutral rack space for use by their own gear and connectivity.
Well, once again, to a large extent, Sprint has done exactly that -- they spun off a new company (creatively named "TowerCo") to own their towers a couple of years ago (it might have been required by the FCC or FTC as a condition of purchasing Nextel). Sprint profits from TowerCo, but doesn't restrict it from leasing space to other carriers (though there's almost certainly a contractual obligation somewhere that guarantees Sprint a price that's no higher than what's charged to anyone else).
From what I've read, it's rare for AT&T and Verizon to use TowerCo's towers, but that's more due to the culture and internal policies of AT&T and Verizon. AT&T and Verizon have a very "Bell" mentality that overwhelmingly disfavors the use of anything they don't control directly.
>And how is this of any use to somebody who doesn't have a CDMA phone. >I'm trying to get rid of the tower-tied-to-one-carrier-or-maybe-two model here.
There's nothing inherently carrier-proprietary about CDMA, it's just the fsck'ed up way Sprint and Verizon implemented it. CDMA has a perfectly good standard for interoperability that's a superset of SIM cards called R-UIM (which itself is an optional subset of USIM). Go to a country like India or China, and CDMA phones are as network-agnostic and interoperable as GSM phones are in Europe. It's only in the US where Sprint refuses to let you use any phone they didn't sell, and Verizon does wacky non-publicly-documented things with their control protocols that cause most unlocked non-Verizon phones activated on Verizon to be dysfunctional in some annoying (if not major) way.
In a very real sense, America's mobile phone industry is like Japan's -- very, very proprietary, mostly because nobody ever had the balls to force them to at least try and not be arbitrarily incompatible. Expecting Sprint phones to be interoperable with AT&T is a pipe dream, but there's no sane reason why a Droid X2 shouldn't work just fine on Sprint, or an Evo 3D shouldn't work perfectly well on Verizon (at least, if you don't care about wimax or LTE right now). And really, there's no long-term reason why LTE and Wimax really need different radio modems. At the raw radio modulation level, LTE and Wimax are almost indistinguishable from each other. The real differences lie with protocol and software (LTE arranges the bits in a way that makes it slightly more power-efficient than Wimax, but the real-world difference is almost academic; it's something that was done because it's a sensible improvement that occurred to someone working on LTE after Wimax was already finalized, but it's not a night-and-day difference).
Moreover, there's no reason a tower has to be tied to one carrier or technology. If Sprint merged with T-Mobile, their newly-deployed tower equipment could be programmed to start providing 1700/2100MHz UMTS (and probably 1900MHz legacy GSM) service at every single existing Sprint tower site within a matter of weeks, because the hardware itself already has the capability of doing it -- Sprint just doesn't have the spectrum or a reason to do it. Legacy GSM can't share channels or timeslots with CDMA voice, UMTS, or EVDO, but they can all coexist on the same tower and cell site without problems. Worst-case, Sprint might need to add a new set of antennas tuned for 1700MHz and 2100MHz.
The problem is that it's hard enough to keep track of all the different Android builds available for *your own* phone, and possibly its close cousins, without even thinking about trying to do it for other brands too. Just look at the forums for Cyanogen. The guys trying to port to to Samsung phones can barely carry on a coherent conversation with the guys who've ported it to HTC phones, because their stock firmware is so architecturally different. You'd think they'd be similar because they're all ARM-based, overwhelmingly use Qualcomm radio chipsets, and all theoretically run Android... but software-wise the differences start at the kernel and device drivers, and just explode from there.
Just to give an idea of the problem's scope, look at the Samsung Galaxy S family. In theory, they were similar cousins. In reality, you couldn't even use an unaltered ROM meant for a Captivate (AT&T) with a Vibrant (T-Mobile) or i9000 (international), let alone an Epic 4G (Sprint) or Fascinate (Verizon). With great effort, you could make a Captivate-origin ROM that worked on a Vibrant, and grab bits and pieces from the Fascinate and stuff it into an Epic4G ROM, but it absolutely wasn't plug-and-play. And these were phones that were supposedly fraternal twins or first cousins. Compare any of them to anything by HTC or Motorola, and you might as well compare an Orangutan to a Kangaroo. I suspect things might be easier in Europe, but in America, even nominally-GSM phones for AT&T and T-Mobile have major differences.
You don't lose SenseUI from *rooting*, you lose SenseUI from replacing its stock ROM with most community Android builds. The main complaint today about most factory ROMs is that there's no graceful way to pick and choose what you want to keep. To a very, very large extent, you can either poke around and rearrange the furniture a bit (leaving most of the original stuff in place), or you can blow it all away and end up with something that often isn't quite as polished or pretty as what you had before.
The main problem is that the Android team largely left it up to manufacturers to implement core stuff like the Dialer app, and never formally defined how a "Dialer" should interact with a "Phonebook" or "Calendar". So what happens is that someone makes a custom ROM, tries tweaking the Dialer, discovers he can't, blows it away and replaces it, then discovers that it can't seamlessly integrate with anything else on the phone because it doesn't know how to interact with the phonebook or calendar. SO... he reverse engineers the phonebook and calendar on HIS phone, gets it to work with his Dialer of choice, then others try to use it and it blows up on their phones because the phonebook and calendar on THEIR phones communicates in a different way than the phonebook and calendar on HIS phone.
THIS is what people really mean when they talk about Android's "fragmented" frameworks -- there's no official standard for how a modular and extensible dialer app should work or interact with the rest of the system, so every new Dialer ends up being specific to a very small specific group of phones, and version upgrades that upgrade the Dialer app end up breaking everything that was based on the old version's reverse-engineered behavior. SenseUI does things one way, Touchwiz does things another, Motoblur does them a third, and AOSP is off in its own world with several other ways for different families of Dialers+phonebooks to interact with each other and the rest of the world.
I believe one of Google's goals for ICS has been to formally define aspects of the "phonebook/contacts/schedule" system and standardize the intents, so that at least going forward manufacturers who properly implement them will have phones that can be incrementally tweaked without having to blow everything away and throw the baby out with the bathwater the way you (mostly) do now.
Slashdot Mirror
The root of Sprint's thrashing problem is the RIL programming that prevents simultaneous wimax and CDMA (EVDO or 1xRTT) data sessions. Now, when a Sprint phone attempts to connect to Wimax, it first terminates any live data session (EVDO or 1xRTT) before it even tries to authenticate to Clear. Likewise, if the wimax connection fails (or is failing), the phone will first fully terminate the wimax session before it even tries to establish a CDMA data connection.
What Sprint really needs to do is give us an option to keep a hot EVDO or 1xRTT data session as a live backup whenever wimax is active, and have the phone establish a fairly conventional multilink VPN with dynamic failover to a VPN server somewhere (favoring wimax, but transparently rerouting data to EVDO or 1xRTT the moment a failure occurred). That way, applications and the outside world would both see a stable IP address, regardless of how ugly things got between the phone and towers.
Technically, this is entirely possible to do with the phones we have RIGHT NOW. In all of Sprint's current phones, the wimax and wi-fi radios are on the same daughtercard (in fact, I think they're the same chip). If it weren't possible to do simultaneous 1xRTT and wimax, it wouldn't be possible to do simultaneous wi-fi and voice calls, either (voice calls and 1xRTT work the same way). It's a shame the radios aren't as hackable by end users as Android itself is. Otherwise, I'd give it 3 months, max, before custom modems started to appear that could use thirdparty VPN services together with simultaneous CDMA + wimax data connections to achieve instant failover. Battery life would probably suck, but at least you'd never find yourself swearing at Connectbot because you forgot to disable 4G before connecting, then at some point the phone spontaneously decided to kill your SSH session so it could try and connect to a new wimax tower it just noticed....
^^^ What he said. When Qualcomm introduced CDMA in the early 90s, it was about as close as you could get to black magic. Engineers from Nokia or Ericsson went on record officially calling it a fraud, and ultimately had to eat their words a couple of years later when they were proven wrong.
The coolest thing about CDMA is that you can literally do things like throw down a low-power cell in the middle of a congested area (like a mall, university, stadium, etc), add it to the network, and watch the congestion just kind of automatically sort itself out and self-optimize the power levels. CDMA networks can grow organically, and in an adhoc manner that would be inconceivable with legacy GSM (not UMTS, because UMTS is CDMA)
>Or if you like to actually own a phone and can swap out SIM cards to change phone numbers and carriers quickly.
CDMA has a perfectly good standard called R-UIM that's a superset of GSM SIM, and an optional subset of USIM, that serves the exact same purpose. Unfortunately, Sprint and Verizon never bothered to support it. Elsewhere in the world, CDMA phones and networks are as mutually interoperable and carrier-agnostic as GSM phones.
Once you've hit the home button and launched the camera app by finding its icon and clicking it. As opposed to the barbaric, uncivilized way Android owners do it... by pressing a (gasp) dedicated camera button ;-)
I don't remember the exact year offhand, I guess it might have been closer to 15 years ago. I just remember that the car was manufactured and bought the final year CFC-using air conditioners were legal, and the repair happened about 5 years later. In retrospect, I probably should have taken the car to get fixed by some random mechanic in his driveway in Hialeah (who could have refilled it with used refrigerant recovered from a junkyard car), but I made the mistake of taking the car to the dealer and got completely raped, just like a few million other Floridians who had no idea at the time how expensive something that was historically a minor band-aid repair job you did to older cars in their final year or so of life before trading them in had just become (the AC in question wasn't completely broken... it had just lost enough refrigerant over the years to eventually start icing up regularly, and really needed only a fairly small amount to top it off).
So, how exactly did a level of CFCs that are now a fraction of a percentage point of any modern-era level -- to the point where they're now banning fucking *asthma inhalers* -- somehow cause the largest "ozone hole" in recorded history? This seems to suggest that we've either gone completely overboard trying to banish every last trace of something useful in an act of complete futility and political masturbation, last year was an epic flaming catastrophe with millions of halon-extinguished fires (since that's pretty much the only significant source of CFCs left), or CFCs don't have nearly as much to do with it as originally speculated, and the "holes" are going to keep occurring regardless of whether or not CFCs continue to exist.
This is not to deny that the phenomenon exists, or that CFCs have no impact, but rather to point out that the worldwide banning of CFCs might have ultimately caused a phenomenon that's going to occur anyway to now be 99% as large as it might otherwise have been instead of 100%, and call into question whether that small improvement was actually worth the expense and misery of its opportunity cost. As someone who got hit with a $1,800 AC repair bill ~9 years ago for something that should have been a $65 refill (because they had to basically tear out and replace the car's entire air conditioner), I have a bit of an axe to grind over this matter. Frankly, I'm not convinced that I've gotten $1,735 worth of added ozone value in return for my investment.
There's a difference between arguing that CFCs shouldn't have been phased out at all, vs arguing that the extreme and expensive measures taken to banish them almost overnight regardless of cost or benefit instead of simply requiring that new equipment be CFC-free going forward were worth it.
> People seem to want symbolic icons that represent the programs they want to run;
> they don't want to look through a long menu and read a bunch of text.
Oh god, no. Please don't remind me about Lotus SmartSuite's Hieroglyphics from the mid-late 90s.
Give me a nice, recognizable icon AND text, so I can recognize the icons I care about frequently, then find the remaining functions without having to play "guess what this is supposed to be symbolic of".
It's still stupid for them to not support 1700MHz uplinks. Even if AT&T *does* buy T-Mobile, it's not like they could instantly absorb all of their customers into AT&T's existing 3G network overnight without making life suck even MORE for existing AT&T customers. If the iPhone 4s phones sold by AT&T supported 1700/2100 (AWS) UMTS, and AT&T closed the deal, they could instantly add T-Mobile's existing 1700/2100 HSPA+ sites to AT&T overnight with little more than a PRL update.
That said, is the iPhone 4s physically INCAPABLE of 1700MHz uplinks, or is it merely not CERTIFIED for them? I find it hard to believe that Apple has somehow managed to buy millions of WCDMA baseband processor chips that are LITERALLY silicon-incapable of 1700MHz operation. Personally, I suspect the iPhone 4s's bare-metal hardware is perfectly capable of 1700MHz AWS, but won't be submitted to the FCC for certification as such unless the AT&T merger falls through.
The main risk to sitting upright is aortic dissection. Apparently, when the body comes to a rapid stop in a forward-facing upright seated position, your internal organs can shift forward before coming to a stop and cause the aorta to tear enough to cause you to die from uncontrollable internal bleeding -- possibly minutes or hours AFTER you've gotten off the plane and are busy celebrating your survival.
Not quite right. South Korea USED to be 100% CDMA, but a couple of years ago the carriers decided to skip EVDO and go with W-CDMA/UMTS instead. The result is that South Korea's cellular network is like Canada's and semi-rural Australia's -- CDMA2000 for legacy voice and 1xRTT data is available everywhere, 3G is strictly UMTS, and some new phones sold there can't do CDMA2000 at all.
Or, if you're a Sprint customer due for an upgrade, buy a Motorola Photon. It does both GSM and CDMA. Just make sure you get the SIM unlock code from Sprint before you leave, and be aware that the SIM is used ONLY for GSM identity (in theory, Sprint could have used it with a USIM card capable of operating as a CDMA R-UIM... but they didn't), so you won't be able to casually use it on any CDMA network in China (except as an expensively-roaming Sprint phone).
If you're trying to make a blanket, universal declaration, yes it's evil. In this specific context, though, you have the database of a worldwide book retailer bought by another one that's more or less equivalent, who plans to use it for more or less the exact same purpose in the same way as Borders.
Now, things might be different if the mailing list were purchased by some religious group who planned to scour it for "unwholesome" purchases to identify "sinners" and make their lives miserable... and that's why there should be a degree of judicial oversight and a legal requirement that any such transfer be approved by the court, and only to a purchaser with the intent and means of using it in the same way for substantially the same purpose as before. Barnes & Noble clearly falls into that category. Amazon mostly does (no actual brick & mortar stores). Wal Mart or Buy.com would be a slippery case. The New Reformed Church of Jesus Christ the Vengeful Redeemer (and its 17 members, not including the psychotic preacher's family) from somewhere in rural Mississippi would be an obvious example of the other extreme. The hard part is drawing the line in a way that makes it easy for someone like B&N to buy it, and impossible for someone like the NRCJCVR to buy it.
A reasonable compromise might be a law that placed the data in escrow, and allowed its purchaser a single opportunity to contact the individuals via email and allow them to either formally opt in and allow their data to be handed over to the purchaser, or do nothing and have their data be permanently destroyed 6 months later. With a further rule that no single database can be sold to more than 3 potential purchasers, and that anyone receiving the first email can explicitly opt out of both the current offer and all remaining ones with a single click.
I can see why some are objecting to this specific transfer for the sake of having a consistent policy, but hysteria over it is just making the broader cause look silly.
AT&T was able to catch iPhone users who tethered because all but one of the iPhone tethering apps hijacked IOS' built-in tethering support, which was trivially easy for AT&T to track due to the way it worked. An app like EasyTether for Android is effectively untraceable by carriers if you use it to establish an IPSEC VPN, because even with deep packet inspection it's indistinguishable from an android app creating the same VPN for its own use. If necessary, we can raise the stakes and use SSH to resurrect SlipKnot to our own servers. Or masquerade it as an RDP back-channel audio stream.
There's a difference between carriers being able to cast a net and catch unsophisticated iPhone users who downloaded some point and click tethering app, and being able to catch people who are hellbent on sliding under the radar and have the technical skill to pull it off.
Pair-bonded IDSL is basically two IDSL lines aggregated as PPP Multilink. It won't do much for a single user using something like RDP or VNC, but for web browsing that involves one request for a page followed by a burst of requests for separate files, it effectively doubles it.
If you're willing to get down & dirty with Covad, pull out the checkbook, and let them work their magic on dry copper pairs leased from the ILEC, they can do even more impressive things and go places where mainstream mass-market ISPs are afraid to tread. I've read of people in rural Tennessee who've gotten 460kbit and more 25 miles from the nearest CO. The catch is, it's not cheap. They're basically building a private ISP for you alone by leasing rackspace at a central office, and pulling off some fairly expensive DSP magic (possibly with repeaters along the way to regenerate the signal) to get you the fastest data possible.
It's not mainstream, because most "rural folks" who aren't wealthy New Yorkers living in a cabin in the mountains because it's cute and cool can't afford to spend several hundred dollars per month for what's basically commercial-grade internet access, but you'd be amazed at what you can get if installation fees in the neighborhood of $1k-5k and monthly service fees between $250 and $500 are tolerable. I think the pair-bonded IDSL route is the most popular solution in Tennessee, because the way phone service is regulated there makes ISDN dirt cheap compared to the rest of the country, so it's not prohibitively expensive to just order more lines and aggregate them in parallel with PPP multilink the way people used to shotgun 56k modems in the late 90s.
> Or how would you recommend depositing a check/cheque to such a bank?
If you don't want to spend significant amounts of money on internet service? With a PC, a scanner, and a dialup modem. Worst-case, you can probably run the Android emulator in wifi mode (not sure whether the emulator can use the PC's webcam, I've never actually tried it).
Well, strictly speaking, they didn't "block" them, they just didn't allow them to be shown in Android Market. They made it non-easy for unsophisticated users, but didn't actually make it *hard* for regular users the way AT&T did.
Now, if they started poisoning DNS to make their domain appear to be invalid, or started to actually intercept and mangle http requests to their web site, that would be much more incrementally-evil and condemnation-worthy. On a scale of 1 to 10:
Filtering from Android Market: 2
DNS poisoning: 7
Http filtering: 8
Disabling installation of non-market apps: 9
Disabling ADB: 10 (I think this is actually forbidden by Google's licensing)
AT&T in General: 17
AT&T's management: off the scale. Alderaan is fucked, and only the FTC can save the rest of us.
If Sprint DOES end up with exclusivity, I can almost guarantee it won't extend beyond December 1. Honestly, I wouldn't expect it to last beyond November 15th. Sprint will probably get to have it to themselves for four weeks while Foxconn builds more for AT&T and Verizon, then Apple throws open the floodgates to make as much money as it can before Christmas. There's no way in HELL Apple would risk losing Christmas sales to people who'd prefer an iPhone, but will take a high-end Android phone if it means having it to play with on Christmas Day.
For the sake of accuracy, the only carrier known to have ever done that in the US is AT&T, and they appear to have quit doing it for new phones going forward from the Infuse, and supposedly are unlocking older phones as they roll out periodic updates over the next few months. Now, whether AT&T will KEEP leaving them unlocked if it loses its fight to buy T-Mobile, and quits trying to publicly pretend that it's non-Evil, is anybody's guess.
> Now you understand why Samsung just hired Steve Kondik, founder of the Cyanogenmod project.
> They need someone like him very badly.
You're absolutely right. Actually, Steve will help Samsung a lot, because for basically the cost of one happy full-time employee, they've effectively outsourced the long-term maintenance of their phones' firmware to dozens to hundreds of enthusiastic, highly-skilled unpaid volunteers (many of whom would be VERY expensive to hire for real as full-time employees). Samsung has NEVER been good about supporting phones with updates after they've shipped. Cyanogen went a long way towards fixing that, but had a problem -- there were hardware issues that just couldn't be easily solved without access to the proprietary bits of source that Samsung couldn't hand out to members of the public. Now that Steve's an employee under NDA, they can give him the keys to the castle and let him freely build flawless Cyanogen-optimized kernels for Samsung's phones, and leave everything else up to the community.
> Except perhaps people who live outside the coverage area of cable and DSL service, where nobody offers
>"high-end internet access at home", where the best options for home Internet are satellite and fixed 3G/4G
Please, enlighten me. Name one Google'able address anywhere in America where there's solid 300kbps+ EVDO coverage from Sprint, but cable internet or pair-bonded IDSL of comparable speed or better is unavailable at any cost. No, 153kbps 1xRTT doesn't count, nor does the ability to mount a directional antenna on the roof and get fixed wireless service through Wimax. I'm talking about literally whipping out an Evo indoors and getting anything better than sub-ISDN speeds if you're lucky. People who practically live in the shadow of Sprint's towers can barely get 4G to work indoors, and you honestly think you're going to have viable 4G service on a PHONE somewhere that's so far out in BFE, you can't even get pair-bonded IDSL or cable internet?
iPhones and Android phones aren't for frugal people who want to spend $12/month to make emergency calls. They're for people who live online 24/7, and would dump satellite TV and their landline phone long before they even contemplated giving up unlimited wireless data and top-tier broadband.
> Providing more price-competitive packages.
Compared to what? Sprint has plenty of warts, but price (at least, for individual customers with no family plan and exactly one phone who'd burn through AT&T and Verizon's caps within a matter of days and are perfectly cool with $69.95 + $10 for 450 minutes of peak airtime, and more or less unlimited everything else) isn't really something I'd classify as one of them.
> Providing better 4G caps than the competition.
Root your phone like everyone else, and the 4g hotspot caps are meaningless. Sprint chose that specific group very carefully -- the nontechnical users who actually go out and PAY for an official hotspot plan are almost exactly the same group who'll try to use their phone as their one and only means of internet access.
Sprint isn't stupid... they know the overwhelming majority of users who root, reflash, and tether for free already have the most expensive cable or DSL internet access they can buy, and use it instead of their Sprint data service when they're at home just because it's faster and works better. To repeat: Sprint really doesn't care about users who tether once in a blue moon so they can get online with their laptop at an airport somewhere while waiting for a plane. Sprint passionately cares about users who try to use tethering as a substitute for real internet service and 21st-century dialup.
Nobody who has high-end internet access at home is going to screw around with torrenting from a tethered phone, because it would be slow, suck, and annoyingly cause most of your incoming calls to end up going straight to voicemail. Likewise, statistically nobody with the means to tether is going to stream lots of HD video, because it's not free -- users who tether for free rip their content from Blu-Ray, convert it to .mkv, copy it to their 32-gig Class 10 microSD cards, and watch it from THERE. Sprint is one of the few carriers who understands that the users who can most easily subvert any controls they try to impose are likely to be the ones who fall towards the lower end of total monthly data use, simply because those users have better ways of getting online anyway.
> Upgrading network capacity.
No arguments there. Sprint definitely has plenty of room to improve in that regard.
^^^ Just to clarify one point that I just thought of... TowerCo owns the land/lease and the tower itself. TowerCo does NOT own the spectrum licenses or operate the actual equipment. It's the equivalent of a colo facility where carriers (including Sprint) rent the equivalent of neutral rack space for use by their own gear and connectivity.
Well, once again, to a large extent, Sprint has done exactly that -- they spun off a new company (creatively named "TowerCo") to own their towers a couple of years ago (it might have been required by the FCC or FTC as a condition of purchasing Nextel). Sprint profits from TowerCo, but doesn't restrict it from leasing space to other carriers (though there's almost certainly a contractual obligation somewhere that guarantees Sprint a price that's no higher than what's charged to anyone else).
From what I've read, it's rare for AT&T and Verizon to use TowerCo's towers, but that's more due to the culture and internal policies of AT&T and Verizon. AT&T and Verizon have a very "Bell" mentality that overwhelmingly disfavors the use of anything they don't control directly.
>And how is this of any use to somebody who doesn't have a CDMA phone.
>I'm trying to get rid of the tower-tied-to-one-carrier-or-maybe-two model here.
There's nothing inherently carrier-proprietary about CDMA, it's just the fsck'ed up way Sprint and Verizon implemented it. CDMA has a perfectly good standard for interoperability that's a superset of SIM cards called R-UIM (which itself is an optional subset of USIM). Go to a country like India or China, and CDMA phones are as network-agnostic and interoperable as GSM phones are in Europe. It's only in the US where Sprint refuses to let you use any phone they didn't sell, and Verizon does wacky non-publicly-documented things with their control protocols that cause most unlocked non-Verizon phones activated on Verizon to be dysfunctional in some annoying (if not major) way.
In a very real sense, America's mobile phone industry is like Japan's -- very, very proprietary, mostly because nobody ever had the balls to force them to at least try and not be arbitrarily incompatible. Expecting Sprint phones to be interoperable with AT&T is a pipe dream, but there's no sane reason why a Droid X2 shouldn't work just fine on Sprint, or an Evo 3D shouldn't work perfectly well on Verizon (at least, if you don't care about wimax or LTE right now). And really, there's no long-term reason why LTE and Wimax really need different radio modems. At the raw radio modulation level, LTE and Wimax are almost indistinguishable from each other. The real differences lie with protocol and software (LTE arranges the bits in a way that makes it slightly more power-efficient than Wimax, but the real-world difference is almost academic; it's something that was done because it's a sensible improvement that occurred to someone working on LTE after Wimax was already finalized, but it's not a night-and-day difference).
Moreover, there's no reason a tower has to be tied to one carrier or technology. If Sprint merged with T-Mobile, their newly-deployed tower equipment could be programmed to start providing 1700/2100MHz UMTS (and probably 1900MHz legacy GSM) service at every single existing Sprint tower site within a matter of weeks, because the hardware itself already has the capability of doing it -- Sprint just doesn't have the spectrum or a reason to do it. Legacy GSM can't share channels or timeslots with CDMA voice, UMTS, or EVDO, but they can all coexist on the same tower and cell site without problems. Worst-case, Sprint might need to add a new set of antennas tuned for 1700MHz and 2100MHz.
The problem is that it's hard enough to keep track of all the different Android builds available for *your own* phone, and possibly its close cousins, without even thinking about trying to do it for other brands too. Just look at the forums for Cyanogen. The guys trying to port to to Samsung phones can barely carry on a coherent conversation with the guys who've ported it to HTC phones, because their stock firmware is so architecturally different. You'd think they'd be similar because they're all ARM-based, overwhelmingly use Qualcomm radio chipsets, and all theoretically run Android... but software-wise the differences start at the kernel and device drivers, and just explode from there.
Just to give an idea of the problem's scope, look at the Samsung Galaxy S family. In theory, they were similar cousins. In reality, you couldn't even use an unaltered ROM meant for a Captivate (AT&T) with a Vibrant (T-Mobile) or i9000 (international), let alone an Epic 4G (Sprint) or Fascinate (Verizon). With great effort, you could make a Captivate-origin ROM that worked on a Vibrant, and grab bits and pieces from the Fascinate and stuff it into an Epic4G ROM, but it absolutely wasn't plug-and-play. And these were phones that were supposedly fraternal twins or first cousins. Compare any of them to anything by HTC or Motorola, and you might as well compare an Orangutan to a Kangaroo. I suspect things might be easier in Europe, but in America, even nominally-GSM phones for AT&T and T-Mobile have major differences.
You don't lose SenseUI from *rooting*, you lose SenseUI from replacing its stock ROM with most community Android builds. The main complaint today about most factory ROMs is that there's no graceful way to pick and choose what you want to keep. To a very, very large extent, you can either poke around and rearrange the furniture a bit (leaving most of the original stuff in place), or you can blow it all away and end up with something that often isn't quite as polished or pretty as what you had before.
The main problem is that the Android team largely left it up to manufacturers to implement core stuff like the Dialer app, and never formally defined how a "Dialer" should interact with a "Phonebook" or "Calendar". So what happens is that someone makes a custom ROM, tries tweaking the Dialer, discovers he can't, blows it away and replaces it, then discovers that it can't seamlessly integrate with anything else on the phone because it doesn't know how to interact with the phonebook or calendar. SO... he reverse engineers the phonebook and calendar on HIS phone, gets it to work with his Dialer of choice, then others try to use it and it blows up on their phones because the phonebook and calendar on THEIR phones communicates in a different way than the phonebook and calendar on HIS phone.
THIS is what people really mean when they talk about Android's "fragmented" frameworks -- there's no official standard for how a modular and extensible dialer app should work or interact with the rest of the system, so every new Dialer ends up being specific to a very small specific group of phones, and version upgrades that upgrade the Dialer app end up breaking everything that was based on the old version's reverse-engineered behavior. SenseUI does things one way, Touchwiz does things another, Motoblur does them a third, and AOSP is off in its own world with several other ways for different families of Dialers+phonebooks to interact with each other and the rest of the world.
I believe one of Google's goals for ICS has been to formally define aspects of the "phonebook/contacts/schedule" system and standardize the intents, so that at least going forward manufacturers who properly implement them will have phones that can be incrementally tweaked without having to blow everything away and throw the baby out with the bathwater the way you (mostly) do now.