> Commercial software that gets pirated is useful by definition... otherwise it would not be pirated
You underestimate the number of packrats with ADD -- users who have dozens and dozens of pirated apps installed... 99% of which they never use, and only keep around because having lots and lots of apps makes them happy. The trick to monetizing these users is to create an app that has two levels of piracy... an "easy" level of cracking that lets them feel like they got the real app, then a much harder level of cracking that suddenly manifests itself at a point when they urgently need the app to work and will *instantly* pull out the credit card and spend a few bucks just to solve their problem *immediately*.
No, I'm not talking about apps that are borderline-malicious. For example, suppose you wrote an app to lay out PC boards (unsuitable for a phone, but this is just for illustration). Officially, you limit it to 2x2 inches and one side, but make it fairly easy to crack and allow nominally unlimited size, two sides, and four layers. HOWEVER... the app knows it's pirated, so it just sits and waits. And waits, until the user goes to export it to a Gerber file for manufacturing. You even allow him to export as many Gerber files as he wants to, until 5 minutes elapse without an export attempt. Then you pull the trigger -- the next time he goes to print a Gerber, make it look like your app has somehow reverted to "lite" mode and needs to be purchased to continue. No, you don't jack up the price at that point... remember, the goal isn't to piss him off and motivate him to go hunting for a crack. The goal is to get him to the point where he's stressed out, racing to meet a deadline, and desperate... then hit him with a reasonable charge that will make the problem go away forever. For this to work, you have to make it blame-free, easy, fair, and (most importantly) *guaranteed to be instant*. If you tell him his order will be processed within 24 hours, you've just lost the sale. He's going to go right back to hunting for a crack so he can fix it *now*, because you just admitted that buying it *won't* solve his problem *immediately*.
Something similar can be applied to games (and, in fact, HAS been applied to games, but in a way intended more to extract revenge on pirates rather than drive sales). A game appears to be cracked and goes along with it, then at some critical moment pretends to have reverted to demo mode in a way that can be instantly restored to full mode upon purchasing an activation code (possibly from within the game itself). Pull out your Visa card in the next 5 minutes, and you can slay the dragon & save the day. Spend too long thinking about it, and you'll be the dragon's lunch & your past 16 hours will have been wasted. The key is to make sure that the only people who even GET to this point are the hardcore gamers who REALLY play it. First, because they're the ones with enough invested in it to pony up the cash to continue. Second, because if you reveal it TOO soon, you'll just motivate some lesser script kiddies to try cracking it for everyone else. Urgency works. Wait until the player is *so* into the game, he won't WANT to waste time cracking it, because that would distract him from playing the game. To pick up more low-hanging fruit, make the game sold and crackable on multiple levels. Free demo, $3-5 base game that's easy to crack, but starts showing ads a few days after cracking, then a $10-25 'advanced' module that pretends to be easy to crack, and just lies in wait for the right moment to make its, *cough* sales pitch *cough*.
It's basic security, really. If someone is hyperfocused on compromising your security, the best thing you can do is to let them think they've won, and get them distracted for a few days. Then, it'll be that much harder for them to continue when you start throwing challenges at them again. They'll have to re-learn things they forgot, and get back into "the zone". You're not going to make a sale to someone who's invested 3 days trying to crack y
> Since I don't believe I can interest you in a support contract for my jump and run game, > I'm going to plaster it with ads as an alternative source of income.
And THERE, you've just hit the nail on the head as to why Google isn't terribly interested in fixing the problems with Android Market's handling of paid apps. Google knows that developers whose paid apps get universally-pirated tend to go with Plan B -- in-app advertisements. 999 times out of a thousand, ads from AdMob, which Google happens to own.
When you sell your app through Google Market, the credit card company gets the biggest chunk, followed by some crumbs for the purchaser's carrier and Google. But when your use of the app drives ad sales, Google reaps 100% of the financial benefit. Ergo, no real motive to make Market work for commercial apps.
Truth be told, Android Market and the way Google runs it sucks *so badly*, I'm genuinely shocked that Valve hasn't partnered with Amazon to launch their own Android app store and bribe the major carriers into giving them equal phone status with Google's. As long as they allowed anyone with a valid credit card to purchase apps (regardless of carrier and/or country, unless the developer/publisher literally bent over backwards to restrict the app's international availability) and didn't completely drop the ball, they'd *own* the market for commercial Android apps within a matter of months.
> The Android market place let's you refund any application within 24 hours of purchase.
There's a hitch -- you only get to do it once per app. So... you trip across something that looks like an interesting app in Market and buy it. It crashes and burns on your phone, or doesn't do what you thought it would, so you get a refund. 9-15 months later, you have a new phone, or they claim to have a substantially improved version available. Unfortunately, if it still crashes on your new phone, or has an intolerable bug/deficiency, you're out of luck. The second purchase is forever. This is a REALLY major problem.
Worst of all, there isn't even any way for the app's publisher to grant an exception, short of writing a refund check himself and personally eating the fees from Google and the credit card processing himself above and beyond the purchase price.
> If you have brand x phone and I have brand y and you have a cool app, will it really work on my phone with a different processor, > screen geometry, camera, sensors, etc?
Basically, yes. For the same reason you can run the same Windows and Linux software regardless of whether your x86 CPU was made by AMD or Intel, and use 3D graphics regardless of whether the video chipset was made by AMD, nVidia, or Intel. Strictly speaking, native code compiled for ARM won't work on an x86-architecture device running Android... but as a practical matter, just about every Android device that matters financially to real-world developers has an ARM processor.
Ditto for frameworks. The "Android Fragmentation" problem isn't due to a need to write one set of programs that work with SenseUI, another set that work with TouchWiz, and another set that work with MotoBlur. It's due to the fact that SenseUI, TouchWiz, and Motoblur keep phone owners shackled to old versions of Android because every new version tends to catastrophically break the manufacturers' proprietary "frameworks" that nothing besides the manufacturer's own apps use, and every major new version of Android has introduced lots of badly-needed basic features that were missing from early versions, so being shackled to an older version of Android really, really sucks. That's why so many Android owners have mixed feelings about SenseUI in particular -- it's very pretty. When it's cutting-edge, it's great. It's shiny, cool, and pretty. But 3 months later, when the next major version of Android gets released that leapfrogs ahead of the version chained down by SenseUI, it's an ugly ball and chain that holds back the rest of Android from progressing until HTC gets SenseUI working with the new version of Android.
They couldn't be using exclusively 1900 for UMTS on any menaingful scale. In 99% of its major east/west-coast urban markets, T-Mobile barely has enough 1900MHz spectrum to handle voice calls and EDGE. It some areas, they don't even have enough spectrum to do EDGE, and all they can do is squeeze GPRS into timeslots next to voice calls (EDGE requires dedicated spectrum, GPRS can share timeslots with voice calls). That's why they bought 1700 and 2100 spectrum in the most recent auction ~4 years ago.
For at least the next 2-5 years, Verizon's LTE network will be about as useful to a visiting European as Sprint's WiMax network -- ie, not useful at all. Verizon's using it ONLY for data, just like Sprint. The truth is, Sprint will have to switch to LTE eventually, if only to avoid being locked in to a single vendor for hardware... but in the meantime, Sprint customers will get to have 4G 6-18 months sooner than their neighbors with Verizon. Verizon's LTE is pretty much the bare-bones minimum they can deploy and still legally call it "LTE". It won't be worse than Sprint's WiMax, but it won't be any better than Sprint's WiMax, either. Two steps forward, three steps back, and one big step diagonally-forward to the right.
Also, there isn't much "4G urgency" among GSM/UMTS networks, in Europe or the US. For Sprint and Verizon, WiMax/LTE is a big deal, because it enables simultaneous circuit-switched voice calls + data, above and beyond any higher speeds or reduced latency. For UMTS networks, LTE is more of an incremental step forward, as opposed to a profound game-changing life-altering upgrade the way it is for CDMA users. AT&T and T-Mobile will scramble to deploy LTE because Sprint and Verizon will hammer away at them in ads and make them look substandard if they don't, but the same urgency doesn't really exist elsewhere in the world.
Wait... you have a webserver running on your phone that's able to take inbound http requests over the mobile network from the internet at large? Who's your carrier?!? AFAIK, every carrier in America (if not the world) effectively firewalls their mobile-phone IP addresses from inbound tcp connections from the outside.
The most credible theory I've seen so far is that the towers and/or Epic4G don't recognize each other as being capable of EVDOrevA, and are falling back to rev0 (which, conveniently, has a reverse data rate of almost exactly 153kbit/sec). I personally doubt Sprint would have done something as stupid as blatantly throttle Epic4G owners down to 150kbit/sec, because they're smart enough to know that Epic4G owners were going to be pulling out the benchmarks and comparing metaphorical penis size with Evo owners from day one, and anything that blatant would have been discovered *instantly*.
The good news is that if it's just a tower-phone identity issue, it's almost certainly something that can be fixed. The bad news is that if it requires tower-config changes, Sprint will probably try to work it into their normal progressive maintenance schedule instead of doing whatever it takes to deploy a potentially-disruptive fix immediately.
Hence, my primary motive for getting this story to Slashdot: harness the power of public relations to light a fire under Sprint's feet and force them to escalate this to a matter of their highest and most urgent priority, instead of plodding along and allowing Epic4G owners to languish at 150k for the next few months.
oops. You're totally right. I'm scratching my head how I ever got that idea. All I can think of is that he might have been clueless about the difference between 3G and GSM, so when he told me he had an imported unlocked phone that only worked in the Toronto area, I went online with my phone, saw it did 1900/2100 UMTS, asked in complete disbelief whether he had 3G, and took his word for it when he said 'yes'.
It does seem kind of strange, though. I can't believe 2100MHz was available in the US, but NOT in Canada... and if the two biggest incumbent carriers (Telus and Bell) both had 1900MHz spectrum and went with UMTS, it almost seems crazy that they *didn't* try to get a hold of 2100MHz spectrum so they *could* have de-facto compatibility with generic international UMTS phones.
> In our backwards little country -- just north of y'all -- the big CDMA vendors have realized > that CDMA sucks from pretty much every standpoint that matters. > Bell and Telus have rolled out nation-wide HSPA networks.
Er, not quite. What happened was that Telus & Bell realized they happened to own 1900MHz spectrum that, combined with 2100MHz spectrum, would enable them to offer direct compatibility with international-standard UMTS phones. So, instead of dropping EVDO alongside 1xRTT (aka "CDMA voice and 150k data"), they dropped UMTS alongside 1xRTT instead. At least, in big cities.
From what I was told by a Canadian friend, until about a year ago, they would ALLOW you to use an imported unlocked phone that was capable of only 1900/2100 UMTS if you insisted, but if you wanted to use it in an area where Telus had no UMTS service (yet), but HAD viable CDMA service, you had to pay the roaming charges yourself. Apparently, they've loosened up this policy a bit so that they'll officially pay your roaming charges... but if more than 50% of your calls end up roaming on UMTS (say, you live in a small town in Saskatchewan), they'll give you a choice between termination, paying the roaming charges yourself, or switching to a phone that can do CDMA.
This wasn't an option at all for Verizon (they're all 850MHz, and 1900MHz is totally owned in America), and wasn't politically an option for Sprint. Sprint has 1900MHz, but would have needed 2100MHz for the downlinks. The FCC had 2100MHz spectrum to sell, but first and foremost wanted to ensure that whomever bought it could use the spectrum to create a viable UMTS network. That drove the price too high to be worth bothering with for Sprint (who really didn't need the 1700MHz spectrum), and kept it affordable for the one network that truly needed it (T-Mobile).
Morally? No. Realistic? Er... um... well, you already know the answer to that.
I have no problem with manufacturer customizations, as long as they give me the phone's root password (I shouldn't *HAVE* to hack my own goddamn phone to get access to it), release the source like they're supposed to, and keep their proprietary driver binaries neatly separated-out from the kernel (Samsung does, HTC doesn't) so I can blow away their junk and replace it with a nice, clean virgin installation of my OS of choice.
That said, as much as I love Android, there's nothing I'd love more than to be able to walk into CompUSA and see retail packages of Windows Phone 7 and WebOS for sale, with a compatibility sticker that includes most of the Samsung, HTC, and Motorola devices in the "PhonePC" aisle...
The solution is quite simple, really. Don't make consumers (#$*@& root their damn phones to upgrade them. Release proper kernel source with proprietary binaries bundled as separate loadable kernel modules, post the manufacturer datasheets to the chips inside the phone along with a reasonable schematic or block diagram showing how everything is wired up, and step out of the way so those same consumers can take matters into their own hands and implement the newest version of Android *themselves*.
If nothing else, the past year has made me vow to never buy an Android phone that can't be rooted and reflashed, even if it means changing carriers if necessary. If you look at the Android phones still stuck with 1.6 today, nearly every last one of them also happens to be a phone whose bootloader remains unconquered by end users (*cough* Motorola). HTC would have loved to drag its feet on the Hero's 2.1 upgrade... but they were getting slaughtered daily online by angry users wanting to know why it was going to take them another 3 months to roll out an upgrade available to users with rooted phones a month earlier. They were literally *forced* to expedite their upgrade strategy, because every day they held back on the official release meant a few thousand more users joining the exodus from the walled garden of officially-blessed upgrades.
No, it's definitely heat and/or atmospheric corrosion. The whole reason why I *have* a CFL in the porch light is so I can leave it on pretty much anytime it's dark outside. South Florida's climate will destroy just about *anything* that's not kept under air conditioning between April and November. My dad has a bolt organizer we bought him ~15 years ago (you know, the ones that have 2 dozen little slide-out drawers with labels, and came with all the differently-sized bolts in little bags) with bolts that have actually started to rust without having ever been anywhere besides in the little plastic drawer tray in the organizer on a shelf in the garage (my personal theory: early morning dew probably condensed in the trays).
Because the fastest phone existing anywhere on earth has approximately the same raw computing power as a 300MHz Pentium II on its best day, ever?
Forget about Flash (a lost cause on anything slower than the equivalent of a 1-GHz Pentium III), just try to do *anything* with the hardware of something like a Nexus One driving a 1920x1080 display in 32-bit color. You'll be lucky if it can render a double-buffered screen update without flickering, let alone do anything like what we've become accustomed to thinking is the norm for even the most ghetto PC you can possibly buy from Wal Mart.
1GHz ARM11 is NOT the equivalent of a 1GHz Pentium M. It's not even the equivalent of a hypothetical 1-GHz Intel Netburst-architecture Celeron. If you cranked an original Cyrix 6x86 up to 1GHz, it might be roughly equivalent to an ARM of the same speed if you found the right benchmark. There's a reason why you don't see ARMs running Windows... or really, even Linux desktop applications. They just don't have the raw horsepower to do it. If you tried to boot Windows 7 on a PC with the ram and horsepower of a Nexus One, you'd be lucky to have it huff and wheeze past the boot logo before collapsing from resource exhaustion.
> does it matter? oh were out of coal and natural gas, time to rape the poor people after they just bought 14$ lightbulbs
I don't know about where you live, but the United States most certainly is "out" of neither. The US has more coal and natural gas than it literally knows what to do with. The problem with coal is that it nasty stuff every inch of the way, from mining to burning. The problem with natural gas is its low energy density, so the only way to viably transport it in bulk long distances in quantities larger than those needed to fuel an occasional barbecue is via pipeline... and US pipeline capacity is grossly inadequate right now. The good news is that new pipelines are under construction... and have been for the past 10 years. The bad news is that they're still about a decade away from making a meaningful dent in winter capacity shortfalls. In the long run, though, if push came to shove, the US has enough of both to last for centuries... at fairly low prices, too.
I personally don't understand the fetish everyone seems to have with LEDs. Joules per lumen, there's almost no meaningful difference at room-lighting quantities between the energy use of CFL and LEDs. Heatsink fans aren't exactly powered by goodwill.
Fluorescent tubes are great when you need lots of relatively diffuse light. LEDs are great when you either need a tiny, tiny bit of light with minimal ceremony or drama, and when you need a fair amount of very, very directional light. They make great backlights, indicator lights, and spotlights. They suck for general room illumination unless you go to ridiculous lengths to try and herd a few hundred of them into simulating the radiation pattern of a normal light bulb. Both have their appropriate uses, and so do incandescent bulbs. I wouldn't use an incandescent bulb for a main light in my house that burns for half the day every day. I most certainly WOULD use an incandescent bulb in a shed where it might burn for 20 minutes per week, and a CFL would be corroded by Florida's climate within a year or two. Humid, salty air does really ugly things to CFL bulbs when you use them in conditions that are semi-indoors, but not climate-controlled (like sheds, garages, etc). I know, because the CFL bulbs in my porch light seem to average 8-14 months of life before they die... incandescent bulbs in the same fixture lasted for years.
Just *try* editing Perl on a server with Connectbot using Swype (or any other soft keyboard). Or SQL. It'll have you in tears, rage, or both. I personally use Graffiti for almost everything, but my next phone is going to be an Epic 4G for one reason: there IS NO soft keyboard that doesn't utterly and completely SUCK with terminal emulation or when typing things that deviate from the content of "normal" text messages. When evil programmers die, they go to hell and spend eternity using Android Scripting Engine to edit Perl using HTC's stock SenseUI soft keyboard (Swype in discrete mode is much better, but it's like comparing "utterly and completely unusable" to "sucks ass" when it comes to typing stuff like that).
> Even if he gave up on waiting he can get Android handsets now with 1Ghz. I would think if Tmobile wants the G2 to be > their flagship phone they would have gone at least on par with the N1.
Based on my own experience with HTC phones, a rooted G2 will run at 1GHz without breaking a sweat... it just won't get acceptable life from a stock battery. Frankly, 90% of the things people bitch about with the current crop of Android phones are due to inadequately-sized stock batteries. If you own a recent-vintage Android phone, do yourself a favor... buy a nice, beefy 3400+mAH battery (roughly double the life of most stock batteries), root it, disable cpu-scaling, and watch your performance problems fade away one by one. Graffiti on a 528MHz Hero (scales down to 200MHz or below) is almost unusable... jack it up to 710MHz and disable speed-scaling, and it magically becomes almost error-free. The culprit wasn't even the 528MHz... it was the fact that HTC programmed the phones to slow down to almost nothing when an input editor is open, on the theory that they're "only" displaying a picture of a keyboard and waiting for a keypress. Violate those assumptions with something trying to do Graffiti, and the whole thing makes a mess. Prevent it from slowing down, and it becomes smooth as glass. Remember, if the power-management takes 100ms to decide you're doing something and speed up the phone, that's 100ms (really, 350-800ms since you can't just jack the speed directly up to max... you have to step it up) it's going to lag when you casually pick up the phone and swipe across the screen.
To repeat: a big battery is the greatest gift you can possibly give yourself and your rooted Android phone. Well, ok... a non-Sandisk class-6 or class-10 microSDHC card for the swapfile is the second (Sandisk is rarely the absolute fastest, because their target market is video professionals and photographers who value stability and shelf-life over balls to the wall random-access small-file performance; the fastest flash comes from companies that overclock the controllers and drive it harder to boost the benchmarks, even if it means the chip will start failing in 3 years instead of 20.)
The Pentium 4 ("Netburst" family) was designed to excel at exactly one thing: provide cheap gigahertz, performance be damned. A 1.2GHz Pentium III Xeon could dance circles around a 3GHz Pentium 4, then blow a final fart at it before sprinting away. It's an example of the kind of disaster you end up with when you allow marketing to dictate engineering (a.k.a. "Marketecture").
The original Pentium M? In a very real sense, it's basically a die-shrink Pentium III Xeon with added power-management features.
The reason why Intel and AMD largely quit focusing on raw clock rate in favor of multi-core chips is because bus speeds higher than ~3GHz (slowly inching higher, but slowly... very, very slowly...) aren't just pointless, they're counterproductive because nothing else in the system can run that fast, so it ends up spending most of its time waiting for the outside world to catch up with it.
From what I've seen, it's kind of like Windows Mobile was... if you spend lots of time learning how to use it and tweaking it, you can make it work very nicely... but a virgin N900, freshly powered up for the first time out of the box from the store, is borderline-dysfunctional as a device for making and receiving voice telephone calls.
Out of box experience matters. It's one reason why Android has done well, while the N900 has not. You can root and customize most Android phones... but you don't *HAVE* to do it just to make the phone work well enough to be tolerable in the first place (at least, not for phones that ship with 2.1 or newer... I won't talk about 1.5...).
If Nokia gave the N900's UI and out of box functionality as much love & attention as they give to their other phones, so it started out with a highly-polished and refined UI that can be customized going forward, it would do a LOT better. It certainly wouldn't hurt if they spent the extra Euro and added support for 850MHz UMTS so it would work on AT&T, Rogers (Canada), and Telestra (Australia), too. When you have an uber-niche device, it makes sense to make it compatible with as many networks as possible, even if it DOES slightly increase the manufacturing cost, because it means you're less likely to get stuck with unsold inventory if it flops in one market.
^^^ Oops, almost forgot... I acknowledged the intense use of CSX tracks between Auburndale and Tampa. However, the tracks between Auburndale Junction and West Palm Beach (via Sebring and Okeechobee) are widely known to be almost useless to CSX for freight. In fact, they lose money on those tracks, because they're forced to keep them maintained to a higher standard than they themselves need for the sake of Amtrak. They're not so completely useless that they really want to lose access to them completely, but their value is low enough that they'd sell them to FDOT for a pittance as long as they were guaranteed the right to still use them late at night (when no passenger trains run) for an occasional freight run.
If FDOT builds Tampa-Orlando HSR so it continues north to downtown Orlando from the outskirts of MCO, and buys trains that can run on normal tracks with freight trains (even if there are few actual freight trains on those tracks in actual daily use), Jacksonville becomes a short train ride from Tampa... HSR from Tampa to Orlando, 80-110mph to Sanford & DeLand, then 110-125mph all the way to the outskirts of Jacksonville (dropping back down to 80 or so for the final 5-10 mile jaunt into downtown Jacksonville). With 100% HSR, Orlando-Jacksonville is even less likely than Orlando-Miami. It would just cost too much relative to the potential market for the service.
The trains even exist... Talgo makes them. They're basically the same idea as Acela, but learning from Acela's mistakes. They can't do 180mph, but they can easily hit 125 maybe 150 on a long, straight run), and 110mph is the fastest you can do on regular tracks with grade crossings (ie, the route from Auburndale to Miami, and Orlando to Jacksonville) anyway.
Er, did you read my entire post? We overlap about 25%, and don't really disagree. My point was that Tampa-Orlando HSR (with "Orlando" specifically meaning "MCO's main terminal, to the complete exclusion of direct service to downtown Orlando") in a vacuum was almost pointless, but as part of a larger system (that doesn't necessarily have to be HSR, even if Tampa-Orlando itself is) with transfer-free service to Miami (and Jacksonville) would be immensely useful.
The biggest single problem with the Tampa-Orlando HSR line is that it was planned in nearly a complete vacuum, with almost total disregard for the way people could actually use it. Part of the problem is that among government planners, encouraging exurban commuting (ie, daily Lakeland-Orlando riders) is taboo, so they ignore the group who could almost single-handedly be its biggest source of reliable daily revenue. They know that today's middle-managers commuting daily from Lakeland to downtown Orlando are tomorrow's CEOs who'll be moving the company HQ to a brand new skyscraper in downtown Lakeland 3 blocks from the train station... and the thought horrifies them.
Likewise, there's the fixation with electric trains. The problem is, electrification (especially somewhere like Florida, where any cat-2 hurricane that brushes by is likely to leave the catenaries out of service for days, or longer) is prohibitively expensive... and FDOT is planning to run ONE TRAIN PER HOUR in each direction. Good god. Nobody ever bothered to tell them apparently that you can ALWAYS leave room to electrify later (when there's enough use to justify it), and live with Diesel for now with few long-term consequences.
Well, I'm not entirely sure of the historical reason, but I do know that the Amtrak situation between L.A. and San Francisco is way worse than it really ought to be, and is almost certainly due to politics. I'm going to guess that 20 years ago, there was a pissing match that came down to
* powerful minority groups (of which Oakland is heavily comprised) demanding service from Washington
* NIMBYs along the other side of the bay wanting nothing to do with Amtrak
* No way to get directly into downtown San Francisco without skipping those minorities and fighting the NIMBYs
Apparently, Union Pacific had something to do with it, too... basically making it so impossible to run passenger trains between LA and the bay area in a reasonably timely manner, it couldn't have succeeded anyway.
Florida's Amtrak situation isn't great (5-6 hours from Tampa or Orlando to Miami, vs ~4-5 to drive), but really, if they could just add 6:30pm trains departing from Tampa & Orlando that arrived in Miami before midnight, and vice-versa, they'd do quite well. The main problem with their current Florida timetable is the fact that you have to take the entire day off from work to ride it, because their only northbound trains leave around 8am and noon, and their southbound trains leave Tampa around noon and Orlando around 1 (the southbound 10:30am train from Orlando heads to Tampa, then continues south to Miami as the noon southbound train from Tampa). Overall, the trip to Orlando or Tampa from Miami is quite nice. In fact, if you spend the extra $54 and get a room, it rocks. Heading south, the room is kind of a waste (they start running out of drinks, food service is limited), but heading north I recommend it 100%. If you have someone with you to split the room cost, it's a no-brainer (the room is $54 extra regardless of whether it's you or you + someone else, and food is included free with it).
> If it was really cost effective some private company would have already built it.
No. At least, not unless it were possible to build today with the kind of free land grants that enabled the original railroad corridors to be constructed 150 years ago.
The fact is, without the authority to condemn land via eminent domain, it would be point blank impossible to build a rail line (or freeway, or even a sidewalk for that matter) of any useful length anywhere in America besides maybe the desert or Alaska -- REGARDLESS of how profitable it might otherwise be once constructed. The moment landowners along the way realized you were building something that needed a continuous path, every last one of them would instantly demand rent-seeking amounts of money for THEIR property. Even if Oprah Winfrey, Warren Buffet, Bill Gates, and Madonna pitched in everything they had to buy the necessary ROW to build a rail line heading north from downtown Miami, they'd collectively be bankrupt before they got to the county line 15 miles north.
Rail lines have an additional disadvantage when it comes to negotiating ROW purchases with individual landowners. Unlike a normal road, which increases the value of land it passes by, a rail line only hurts the property values of adjacent land unless there happens to literally be a station nearby. When stations are 25 miles apart, good luck convincing a landowner 15 miles away that just about anything you care to offer is worth considering, especially if the rail line's construction will effectively cut off access to property on the other side.
Slashdot Mirror
> Commercial software that gets pirated is useful by definition... otherwise it would not be pirated
You underestimate the number of packrats with ADD -- users who have dozens and dozens of pirated apps installed... 99% of which they never use, and only keep around because having lots and lots of apps makes them happy. The trick to monetizing these users is to create an app that has two levels of piracy... an "easy" level of cracking that lets them feel like they got the real app, then a much harder level of cracking that suddenly manifests itself at a point when they urgently need the app to work and will *instantly* pull out the credit card and spend a few bucks just to solve their problem *immediately*.
No, I'm not talking about apps that are borderline-malicious. For example, suppose you wrote an app to lay out PC boards (unsuitable for a phone, but this is just for illustration). Officially, you limit it to 2x2 inches and one side, but make it fairly easy to crack and allow nominally unlimited size, two sides, and four layers. HOWEVER... the app knows it's pirated, so it just sits and waits. And waits, until the user goes to export it to a Gerber file for manufacturing. You even allow him to export as many Gerber files as he wants to, until 5 minutes elapse without an export attempt. Then you pull the trigger -- the next time he goes to print a Gerber, make it look like your app has somehow reverted to "lite" mode and needs to be purchased to continue. No, you don't jack up the price at that point... remember, the goal isn't to piss him off and motivate him to go hunting for a crack. The goal is to get him to the point where he's stressed out, racing to meet a deadline, and desperate... then hit him with a reasonable charge that will make the problem go away forever. For this to work, you have to make it blame-free, easy, fair, and (most importantly) *guaranteed to be instant*. If you tell him his order will be processed within 24 hours, you've just lost the sale. He's going to go right back to hunting for a crack so he can fix it *now*, because you just admitted that buying it *won't* solve his problem *immediately*.
Something similar can be applied to games (and, in fact, HAS been applied to games, but in a way intended more to extract revenge on pirates rather than drive sales). A game appears to be cracked and goes along with it, then at some critical moment pretends to have reverted to demo mode in a way that can be instantly restored to full mode upon purchasing an activation code (possibly from within the game itself). Pull out your Visa card in the next 5 minutes, and you can slay the dragon & save the day. Spend too long thinking about it, and you'll be the dragon's lunch & your past 16 hours will have been wasted. The key is to make sure that the only people who even GET to this point are the hardcore gamers who REALLY play it. First, because they're the ones with enough invested in it to pony up the cash to continue. Second, because if you reveal it TOO soon, you'll just motivate some lesser script kiddies to try cracking it for everyone else. Urgency works. Wait until the player is *so* into the game, he won't WANT to waste time cracking it, because that would distract him from playing the game. To pick up more low-hanging fruit, make the game sold and crackable on multiple levels. Free demo, $3-5 base game that's easy to crack, but starts showing ads a few days after cracking, then a $10-25 'advanced' module that pretends to be easy to crack, and just lies in wait for the right moment to make its, *cough* sales pitch *cough*.
It's basic security, really. If someone is hyperfocused on compromising your security, the best thing you can do is to let them think they've won, and get them distracted for a few days. Then, it'll be that much harder for them to continue when you start throwing challenges at them again. They'll have to re-learn things they forgot, and get back into "the zone". You're not going to make a sale to someone who's invested 3 days trying to crack y
> Since I don't believe I can interest you in a support contract for my jump and run game,
> I'm going to plaster it with ads as an alternative source of income.
And THERE, you've just hit the nail on the head as to why Google isn't terribly interested in fixing the problems with Android Market's handling of paid apps. Google knows that developers whose paid apps get universally-pirated tend to go with Plan B -- in-app advertisements. 999 times out of a thousand, ads from AdMob, which Google happens to own.
When you sell your app through Google Market, the credit card company gets the biggest chunk, followed by some crumbs for the purchaser's carrier and Google. But when your use of the app drives ad sales, Google reaps 100% of the financial benefit. Ergo, no real motive to make Market work for commercial apps.
Truth be told, Android Market and the way Google runs it sucks *so badly*, I'm genuinely shocked that Valve hasn't partnered with Amazon to launch their own Android app store and bribe the major carriers into giving them equal phone status with Google's. As long as they allowed anyone with a valid credit card to purchase apps (regardless of carrier and/or country, unless the developer/publisher literally bent over backwards to restrict the app's international availability) and didn't completely drop the ball, they'd *own* the market for commercial Android apps within a matter of months.
> The Android market place let's you refund any application within 24 hours of purchase.
There's a hitch -- you only get to do it once per app. So... you trip across something that looks like an interesting app in Market and buy it. It crashes and burns on your phone, or doesn't do what you thought it would, so you get a refund. 9-15 months later, you have a new phone, or they claim to have a substantially improved version available. Unfortunately, if it still crashes on your new phone, or has an intolerable bug/deficiency, you're out of luck. The second purchase is forever. This is a REALLY major problem.
Worst of all, there isn't even any way for the app's publisher to grant an exception, short of writing a refund check himself and personally eating the fees from Google and the credit card processing himself above and beyond the purchase price.
> If you have brand x phone and I have brand y and you have a cool app, will it really work on my phone with a different processor,
> screen geometry, camera, sensors, etc?
Basically, yes. For the same reason you can run the same Windows and Linux software regardless of whether your x86 CPU was made by AMD or Intel, and use 3D graphics regardless of whether the video chipset was made by AMD, nVidia, or Intel. Strictly speaking, native code compiled for ARM won't work on an x86-architecture device running Android... but as a practical matter, just about every Android device that matters financially to real-world developers has an ARM processor.
Ditto for frameworks. The "Android Fragmentation" problem isn't due to a need to write one set of programs that work with SenseUI, another set that work with TouchWiz, and another set that work with MotoBlur. It's due to the fact that SenseUI, TouchWiz, and Motoblur keep phone owners shackled to old versions of Android because every new version tends to catastrophically break the manufacturers' proprietary "frameworks" that nothing besides the manufacturer's own apps use, and every major new version of Android has introduced lots of badly-needed basic features that were missing from early versions, so being shackled to an older version of Android really, really sucks. That's why so many Android owners have mixed feelings about SenseUI in particular -- it's very pretty. When it's cutting-edge, it's great. It's shiny, cool, and pretty. But 3 months later, when the next major version of Android gets released that leapfrogs ahead of the version chained down by SenseUI, it's an ugly ball and chain that holds back the rest of Android from progressing until HTC gets SenseUI working with the new version of Android.
They couldn't be using exclusively 1900 for UMTS on any menaingful scale. In 99% of its major east/west-coast urban markets, T-Mobile barely has enough 1900MHz spectrum to handle voice calls and EDGE. It some areas, they don't even have enough spectrum to do EDGE, and all they can do is squeeze GPRS into timeslots next to voice calls (EDGE requires dedicated spectrum, GPRS can share timeslots with voice calls). That's why they bought 1700 and 2100 spectrum in the most recent auction ~4 years ago.
Who's your carrier?
For at least the next 2-5 years, Verizon's LTE network will be about as useful to a visiting European as Sprint's WiMax network -- ie, not useful at all. Verizon's using it ONLY for data, just like Sprint. The truth is, Sprint will have to switch to LTE eventually, if only to avoid being locked in to a single vendor for hardware... but in the meantime, Sprint customers will get to have 4G 6-18 months sooner than their neighbors with Verizon. Verizon's LTE is pretty much the bare-bones minimum they can deploy and still legally call it "LTE". It won't be worse than Sprint's WiMax, but it won't be any better than Sprint's WiMax, either. Two steps forward, three steps back, and one big step diagonally-forward to the right.
Also, there isn't much "4G urgency" among GSM/UMTS networks, in Europe or the US. For Sprint and Verizon, WiMax/LTE is a big deal, because it enables simultaneous circuit-switched voice calls + data, above and beyond any higher speeds or reduced latency. For UMTS networks, LTE is more of an incremental step forward, as opposed to a profound game-changing life-altering upgrade the way it is for CDMA users. AT&T and T-Mobile will scramble to deploy LTE because Sprint and Verizon will hammer away at them in ads and make them look substandard if they don't, but the same urgency doesn't really exist elsewhere in the world.
Wait... you have a webserver running on your phone that's able to take inbound http requests over the mobile network from the internet at large? Who's your carrier?!? AFAIK, every carrier in America (if not the world) effectively firewalls their mobile-phone IP addresses from inbound tcp connections from the outside.
The most credible theory I've seen so far is that the towers and/or Epic4G don't recognize each other as being capable of EVDOrevA, and are falling back to rev0 (which, conveniently, has a reverse data rate of almost exactly 153kbit/sec). I personally doubt Sprint would have done something as stupid as blatantly throttle Epic4G owners down to 150kbit/sec, because they're smart enough to know that Epic4G owners were going to be pulling out the benchmarks and comparing metaphorical penis size with Evo owners from day one, and anything that blatant would have been discovered *instantly*.
The good news is that if it's just a tower-phone identity issue, it's almost certainly something that can be fixed. The bad news is that if it requires tower-config changes, Sprint will probably try to work it into their normal progressive maintenance schedule instead of doing whatever it takes to deploy a potentially-disruptive fix immediately.
Hence, my primary motive for getting this story to Slashdot: harness the power of public relations to light a fire under Sprint's feet and force them to escalate this to a matter of their highest and most urgent priority, instead of plodding along and allowing Epic4G owners to languish at 150k for the next few months.
> 2100MHz is not available in either the US or Canada.
Er, I can't speak authoritatively for Canada, but a chunk of 2100MHz most certainly was freed up by the FCC and sold to T-Mobile ~4 years ago.
oops. You're totally right. I'm scratching my head how I ever got that idea. All I can think of is that he might have been clueless about the difference between 3G and GSM, so when he told me he had an imported unlocked phone that only worked in the Toronto area, I went online with my phone, saw it did 1900/2100 UMTS, asked in complete disbelief whether he had 3G, and took his word for it when he said 'yes'.
It does seem kind of strange, though. I can't believe 2100MHz was available in the US, but NOT in Canada... and if the two biggest incumbent carriers (Telus and Bell) both had 1900MHz spectrum and went with UMTS, it almost seems crazy that they *didn't* try to get a hold of 2100MHz spectrum so they *could* have de-facto compatibility with generic international UMTS phones.
> In our backwards little country -- just north of y'all -- the big CDMA vendors have realized
> that CDMA sucks from pretty much every standpoint that matters.
> Bell and Telus have rolled out nation-wide HSPA networks.
Er, not quite. What happened was that Telus & Bell realized they happened to own 1900MHz spectrum that, combined with 2100MHz spectrum, would enable them to offer direct compatibility with international-standard UMTS phones. So, instead of dropping EVDO alongside 1xRTT (aka "CDMA voice and 150k data"), they dropped UMTS alongside 1xRTT instead. At least, in big cities.
From what I was told by a Canadian friend, until about a year ago, they would ALLOW you to use an imported unlocked phone that was capable of only 1900/2100 UMTS if you insisted, but if you wanted to use it in an area where Telus had no UMTS service (yet), but HAD viable CDMA service, you had to pay the roaming charges yourself. Apparently, they've loosened up this policy a bit so that they'll officially pay your roaming charges... but if more than 50% of your calls end up roaming on UMTS (say, you live in a small town in Saskatchewan), they'll give you a choice between termination, paying the roaming charges yourself, or switching to a phone that can do CDMA.
This wasn't an option at all for Verizon (they're all 850MHz, and 1900MHz is totally owned in America), and wasn't politically an option for Sprint. Sprint has 1900MHz, but would have needed 2100MHz for the downlinks. The FCC had 2100MHz spectrum to sell, but first and foremost wanted to ensure that whomever bought it could use the spectrum to create a viable UMTS network. That drove the price too high to be worth bothering with for Sprint (who really didn't need the 1700MHz spectrum), and kept it affordable for the one network that truly needed it (T-Mobile).
> Is this too much to ask?
Morally? No. Realistic? Er... um... well, you already know the answer to that.
I have no problem with manufacturer customizations, as long as they give me the phone's root password (I shouldn't *HAVE* to hack my own goddamn phone to get access to it), release the source like they're supposed to, and keep their proprietary driver binaries neatly separated-out from the kernel (Samsung does, HTC doesn't) so I can blow away their junk and replace it with a nice, clean virgin installation of my OS of choice.
That said, as much as I love Android, there's nothing I'd love more than to be able to walk into CompUSA and see retail packages of Windows Phone 7 and WebOS for sale, with a compatibility sticker that includes most of the Samsung, HTC, and Motorola devices in the "PhonePC" aisle...
The solution is quite simple, really. Don't make consumers (#$*@& root their damn phones to upgrade them. Release proper kernel source with proprietary binaries bundled as separate loadable kernel modules, post the manufacturer datasheets to the chips inside the phone along with a reasonable schematic or block diagram showing how everything is wired up, and step out of the way so those same consumers can take matters into their own hands and implement the newest version of Android *themselves*.
If nothing else, the past year has made me vow to never buy an Android phone that can't be rooted and reflashed, even if it means changing carriers if necessary. If you look at the Android phones still stuck with 1.6 today, nearly every last one of them also happens to be a phone whose bootloader remains unconquered by end users (*cough* Motorola). HTC would have loved to drag its feet on the Hero's 2.1 upgrade... but they were getting slaughtered daily online by angry users wanting to know why it was going to take them another 3 months to roll out an upgrade available to users with rooted phones a month earlier. They were literally *forced* to expedite their upgrade strategy, because every day they held back on the official release meant a few thousand more users joining the exodus from the walled garden of officially-blessed upgrades.
No, it's definitely heat and/or atmospheric corrosion. The whole reason why I *have* a CFL in the porch light is so I can leave it on pretty much anytime it's dark outside. South Florida's climate will destroy just about *anything* that's not kept under air conditioning between April and November. My dad has a bolt organizer we bought him ~15 years ago (you know, the ones that have 2 dozen little slide-out drawers with labels, and came with all the differently-sized bolts in little bags) with bolts that have actually started to rust without having ever been anywhere besides in the little plastic drawer tray in the organizer on a shelf in the garage (my personal theory: early morning dew probably condensed in the trays).
Because the fastest phone existing anywhere on earth has approximately the same raw computing power as a 300MHz Pentium II on its best day, ever?
Forget about Flash (a lost cause on anything slower than the equivalent of a 1-GHz Pentium III), just try to do *anything* with the hardware of something like a Nexus One driving a 1920x1080 display in 32-bit color. You'll be lucky if it can render a double-buffered screen update without flickering, let alone do anything like what we've become accustomed to thinking is the norm for even the most ghetto PC you can possibly buy from Wal Mart.
1GHz ARM11 is NOT the equivalent of a 1GHz Pentium M. It's not even the equivalent of a hypothetical 1-GHz Intel Netburst-architecture Celeron. If you cranked an original Cyrix 6x86 up to 1GHz, it might be roughly equivalent to an ARM of the same speed if you found the right benchmark. There's a reason why you don't see ARMs running Windows... or really, even Linux desktop applications. They just don't have the raw horsepower to do it. If you tried to boot Windows 7 on a PC with the ram and horsepower of a Nexus One, you'd be lucky to have it huff and wheeze past the boot logo before collapsing from resource exhaustion.
> does it matter? oh were out of coal and natural gas, time to rape the poor people after they just bought 14$ lightbulbs
I don't know about where you live, but the United States most certainly is "out" of neither. The US has more coal and natural gas than it literally knows what to do with. The problem with coal is that it nasty stuff every inch of the way, from mining to burning. The problem with natural gas is its low energy density, so the only way to viably transport it in bulk long distances in quantities larger than those needed to fuel an occasional barbecue is via pipeline... and US pipeline capacity is grossly inadequate right now. The good news is that new pipelines are under construction... and have been for the past 10 years. The bad news is that they're still about a decade away from making a meaningful dent in winter capacity shortfalls. In the long run, though, if push came to shove, the US has enough of both to last for centuries... at fairly low prices, too.
I personally don't understand the fetish everyone seems to have with LEDs. Joules per lumen, there's almost no meaningful difference at room-lighting quantities between the energy use of CFL and LEDs. Heatsink fans aren't exactly powered by goodwill.
Fluorescent tubes are great when you need lots of relatively diffuse light. LEDs are great when you either need a tiny, tiny bit of light with minimal ceremony or drama, and when you need a fair amount of very, very directional light. They make great backlights, indicator lights, and spotlights. They suck for general room illumination unless you go to ridiculous lengths to try and herd a few hundred of them into simulating the radiation pattern of a normal light bulb. Both have their appropriate uses, and so do incandescent bulbs. I wouldn't use an incandescent bulb for a main light in my house that burns for half the day every day. I most certainly WOULD use an incandescent bulb in a shed where it might burn for 20 minutes per week, and a CFL would be corroded by Florida's climate within a year or two. Humid, salty air does really ugly things to CFL bulbs when you use them in conditions that are semi-indoors, but not climate-controlled (like sheds, garages, etc). I know, because the CFL bulbs in my porch light seem to average 8-14 months of life before they die... incandescent bulbs in the same fixture lasted for years.
Just *try* editing Perl on a server with Connectbot using Swype (or any other soft keyboard). Or SQL. It'll have you in tears, rage, or both. I personally use Graffiti for almost everything, but my next phone is going to be an Epic 4G for one reason: there IS NO soft keyboard that doesn't utterly and completely SUCK with terminal emulation or when typing things that deviate from the content of "normal" text messages. When evil programmers die, they go to hell and spend eternity using Android Scripting Engine to edit Perl using HTC's stock SenseUI soft keyboard (Swype in discrete mode is much better, but it's like comparing "utterly and completely unusable" to "sucks ass" when it comes to typing stuff like that).
> Even if he gave up on waiting he can get Android handsets now with 1Ghz. I would think if Tmobile wants the G2 to be
> their flagship phone they would have gone at least on par with the N1.
Based on my own experience with HTC phones, a rooted G2 will run at 1GHz without breaking a sweat... it just won't get acceptable life from a stock battery. Frankly, 90% of the things people bitch about with the current crop of Android phones are due to inadequately-sized stock batteries. If you own a recent-vintage Android phone, do yourself a favor... buy a nice, beefy 3400+mAH battery (roughly double the life of most stock batteries), root it, disable cpu-scaling, and watch your performance problems fade away one by one. Graffiti on a 528MHz Hero (scales down to 200MHz or below) is almost unusable... jack it up to 710MHz and disable speed-scaling, and it magically becomes almost error-free. The culprit wasn't even the 528MHz... it was the fact that HTC programmed the phones to slow down to almost nothing when an input editor is open, on the theory that they're "only" displaying a picture of a keyboard and waiting for a keypress. Violate those assumptions with something trying to do Graffiti, and the whole thing makes a mess. Prevent it from slowing down, and it becomes smooth as glass. Remember, if the power-management takes 100ms to decide you're doing something and speed up the phone, that's 100ms (really, 350-800ms since you can't just jack the speed directly up to max... you have to step it up) it's going to lag when you casually pick up the phone and swipe across the screen.
To repeat: a big battery is the greatest gift you can possibly give yourself and your rooted Android phone. Well, ok... a non-Sandisk class-6 or class-10 microSDHC card for the swapfile is the second (Sandisk is rarely the absolute fastest, because their target market is video professionals and photographers who value stability and shelf-life over balls to the wall random-access small-file performance; the fastest flash comes from companies that overclock the controllers and drive it harder to boost the benchmarks, even if it means the chip will start failing in 3 years instead of 20.)
The Pentium 4 ("Netburst" family) was designed to excel at exactly one thing: provide cheap gigahertz, performance be damned. A 1.2GHz Pentium III Xeon could dance circles around a 3GHz Pentium 4, then blow a final fart at it before sprinting away. It's an example of the kind of disaster you end up with when you allow marketing to dictate engineering (a.k.a. "Marketecture").
The original Pentium M? In a very real sense, it's basically a die-shrink Pentium III Xeon with added power-management features.
The reason why Intel and AMD largely quit focusing on raw clock rate in favor of multi-core chips is because bus speeds higher than ~3GHz (slowly inching higher, but slowly... very, very slowly...) aren't just pointless, they're counterproductive because nothing else in the system can run that fast, so it ends up spending most of its time waiting for the outside world to catch up with it.
From what I've seen, it's kind of like Windows Mobile was... if you spend lots of time learning how to use it and tweaking it, you can make it work very nicely... but a virgin N900, freshly powered up for the first time out of the box from the store, is borderline-dysfunctional as a device for making and receiving voice telephone calls.
Out of box experience matters. It's one reason why Android has done well, while the N900 has not. You can root and customize most Android phones... but you don't *HAVE* to do it just to make the phone work well enough to be tolerable in the first place (at least, not for phones that ship with 2.1 or newer... I won't talk about 1.5...).
If Nokia gave the N900's UI and out of box functionality as much love & attention as they give to their other phones, so it started out with a highly-polished and refined UI that can be customized going forward, it would do a LOT better. It certainly wouldn't hurt if they spent the extra Euro and added support for 850MHz UMTS so it would work on AT&T, Rogers (Canada), and Telestra (Australia), too. When you have an uber-niche device, it makes sense to make it compatible with as many networks as possible, even if it DOES slightly increase the manufacturing cost, because it means you're less likely to get stuck with unsold inventory if it flops in one market.
^^^ Oops, almost forgot... I acknowledged the intense use of CSX tracks between Auburndale and Tampa. However, the tracks between Auburndale Junction and West Palm Beach (via Sebring and Okeechobee) are widely known to be almost useless to CSX for freight. In fact, they lose money on those tracks, because they're forced to keep them maintained to a higher standard than they themselves need for the sake of Amtrak. They're not so completely useless that they really want to lose access to them completely, but their value is low enough that they'd sell them to FDOT for a pittance as long as they were guaranteed the right to still use them late at night (when no passenger trains run) for an occasional freight run.
If FDOT builds Tampa-Orlando HSR so it continues north to downtown Orlando from the outskirts of MCO, and buys trains that can run on normal tracks with freight trains (even if there are few actual freight trains on those tracks in actual daily use), Jacksonville becomes a short train ride from Tampa... HSR from Tampa to Orlando, 80-110mph to Sanford & DeLand, then 110-125mph all the way to the outskirts of Jacksonville (dropping back down to 80 or so for the final 5-10 mile jaunt into downtown Jacksonville). With 100% HSR, Orlando-Jacksonville is even less likely than Orlando-Miami. It would just cost too much relative to the potential market for the service.
The trains even exist... Talgo makes them. They're basically the same idea as Acela, but learning from Acela's mistakes. They can't do 180mph, but they can easily hit 125 maybe 150 on a long, straight run), and 110mph is the fastest you can do on regular tracks with grade crossings (ie, the route from Auburndale to Miami, and Orlando to Jacksonville) anyway.
Er, did you read my entire post? We overlap about 25%, and don't really disagree. My point was that Tampa-Orlando HSR (with "Orlando" specifically meaning "MCO's main terminal, to the complete exclusion of direct service to downtown Orlando") in a vacuum was almost pointless, but as part of a larger system (that doesn't necessarily have to be HSR, even if Tampa-Orlando itself is) with transfer-free service to Miami (and Jacksonville) would be immensely useful.
The biggest single problem with the Tampa-Orlando HSR line is that it was planned in nearly a complete vacuum, with almost total disregard for the way people could actually use it. Part of the problem is that among government planners, encouraging exurban commuting (ie, daily Lakeland-Orlando riders) is taboo, so they ignore the group who could almost single-handedly be its biggest source of reliable daily revenue. They know that today's middle-managers commuting daily from Lakeland to downtown Orlando are tomorrow's CEOs who'll be moving the company HQ to a brand new skyscraper in downtown Lakeland 3 blocks from the train station... and the thought horrifies them.
Likewise, there's the fixation with electric trains. The problem is, electrification (especially somewhere like Florida, where any cat-2 hurricane that brushes by is likely to leave the catenaries out of service for days, or longer) is prohibitively expensive... and FDOT is planning to run ONE TRAIN PER HOUR in each direction. Good god. Nobody ever bothered to tell them apparently that you can ALWAYS leave room to electrify later (when there's enough use to justify it), and live with Diesel for now with few long-term consequences.
Well, I'm not entirely sure of the historical reason, but I do know that the Amtrak situation between L.A. and San Francisco is way worse than it really ought to be, and is almost certainly due to politics. I'm going to guess that 20 years ago, there was a pissing match that came down to
* powerful minority groups (of which Oakland is heavily comprised) demanding service from Washington
* NIMBYs along the other side of the bay wanting nothing to do with Amtrak
* No way to get directly into downtown San Francisco without skipping those minorities and fighting the NIMBYs
Apparently, Union Pacific had something to do with it, too... basically making it so impossible to run passenger trains between LA and the bay area in a reasonably timely manner, it couldn't have succeeded anyway.
Florida's Amtrak situation isn't great (5-6 hours from Tampa or Orlando to Miami, vs ~4-5 to drive), but really, if they could just add 6:30pm trains departing from Tampa & Orlando that arrived in Miami before midnight, and vice-versa, they'd do quite well. The main problem with their current Florida timetable is the fact that you have to take the entire day off from work to ride it, because their only northbound trains leave around 8am and noon, and their southbound trains leave Tampa around noon and Orlando around 1 (the southbound 10:30am train from Orlando heads to Tampa, then continues south to Miami as the noon southbound train from Tampa). Overall, the trip to Orlando or Tampa from Miami is quite nice. In fact, if you spend the extra $54 and get a room, it rocks. Heading south, the room is kind of a waste (they start running out of drinks, food service is limited), but heading north I recommend it 100%. If you have someone with you to split the room cost, it's a no-brainer (the room is $54 extra regardless of whether it's you or you + someone else, and food is included free with it).
> If it was really cost effective some private company would have already built it.
No. At least, not unless it were possible to build today with the kind of free land grants that enabled the original railroad corridors to be constructed 150 years ago.
The fact is, without the authority to condemn land via eminent domain, it would be point blank impossible to build a rail line (or freeway, or even a sidewalk for that matter) of any useful length anywhere in America besides maybe the desert or Alaska -- REGARDLESS of how profitable it might otherwise be once constructed. The moment landowners along the way realized you were building something that needed a continuous path, every last one of them would instantly demand rent-seeking amounts of money for THEIR property. Even if Oprah Winfrey, Warren Buffet, Bill Gates, and Madonna pitched in everything they had to buy the necessary ROW to build a rail line heading north from downtown Miami, they'd collectively be bankrupt before they got to the county line 15 miles north.
Rail lines have an additional disadvantage when it comes to negotiating ROW purchases with individual landowners. Unlike a normal road, which increases the value of land it passes by, a rail line only hurts the property values of adjacent land unless there happens to literally be a station nearby. When stations are 25 miles apart, good luck convincing a landowner 15 miles away that just about anything you care to offer is worth considering, especially if the rail line's construction will effectively cut off access to property on the other side.