Are you serious? I've accidentally "liked", flagged, thumbed-up/down, and other actions while browsing sites with my Android phones over the years, I think developers should be taken out and SHOT for allowing actions with visible, public, and lasting consequences to EVER get triggered by a single touch. They should either require separate confirmation, or give you at least a few seconds to click "ohshit, no, I didn't mean that!!!" to retract/undo them before anyone ever finds out.
PhpBB and other web forums, along with Facebook, are the absolute WORST in this regard. They send out email notifications instantly, so even if you try to retract it, they're going to know you did it anyway.
It should be absolutely, unambiguously legal to photograph a ballot that's in your possession, and publish it however you see fit without restriction.
HOWEVER... any ballot that gets photographed should be deemed "tainted", destroyed (or unambiguously rendered uncountable), and replaced with a fresh ballot (possibly after informing the errant voter that photographed ballots may not be cast, nor may ballots with identifying marks that could be used to identify the ballot as "theirs" in a recount(*). Otherwise, friends, family members, unions, employers, or anyone else could try and coerce members into voting "correctly" by pressuring them to prove that they voted the "right way".
There's a fine line when it comes to appropriate enforcement. It would be wrong to punish somebody who tries to trick officials into accepting his or her photographed ballot... you can only do so much to save someone from themselves... but it WOULD be appropriate to treat it like a "CAPTCHA" situation to prevent someone from attempting to launch a denial of service attack:
The first time someone is caught photographing his or her ballot, they're informed that they have the right to photograph their ballot, but the ballot itself is now legally tainted and can not be cast. It must be destroyed, and they can vote again from a fresh ballot. They'll be informed that they can photograph at their leisure, but would be asked to show consideration for others waiting in line and suggested that they continue their photography at a location where it won't interfere with other voters waiting in line, like the table set up "over there, in the corner".
If they do it a secnod time, they'll be required to watch a short (5-10 minute) video about the importance of secret ballots in a democracy, and asked to place all photography devices in a tamper-evident bag prior to receiving their new ballot. They're told that if the bag is tampered with, or they're observed attempting to photograph their ballot with yet another device, they'll be subjected to additional requirements before they'll be allowed to make their third attempt and beyond.
If they do it a third time, they'll be required to watch the video, wait in line, watch the video again, seal their photography devices in a tamper-evident bag, and write a paragraph expressing their understanding of the rules and their reason. Then vote again on a clean ballot.
If they get caught again, on attempts 4 and beyond, they'll have to watch the video and wait in line once time for each violation beyond the third. So, attempt #4 means watch, wait, watch, wait, watch, write, and vote. Attempt #5 means watch, wait, watch, wait, watch, wait, watch, write, vote. And so on. If they're on attempt #3 or beyond when the polls officially close and the ballot boxes are sealed, and they're stuck waiting, they're out of luck. Essentially, everyone gets one free attempt to photograph and do-over without risk of punishment (beyond having to fill out the ballot again), and you have to actively screw up one more time beyond that before you might end up not being able to vote at all. A hard limit is a serious matter, but if you have someone who's actively gone through the cycle more than twice, it's safe to assume at that point that they're so completely fucked in the head (or coerced, or hellbent on making a political statement by the act itself, or all the above) that it's just a waste of everyone's time to accommodate them further at that point.
This is something that's only going to become a bigger and bigger problem over time. It's also the #1 reason why we can never have internet voting -- there's no way to guarantee that somebody isn't standing over the voter's shoulder, ready to reward them for voting "properly", or punish (real, or implied) them if they don't cooperate and vote the right way.
(*) I'd go a step further, and require that ballots be designed in a way that isolates them into clusters of comparable races, with a maximum of 3 or 4 races per physical ballot-sheet being r
The overnight & 2-day boxes don't go anywhere *near* the 'economy' trucks. Think of the economy trucks as a bulk shipping crate with wheels, whose workflow is completely independent of faster methods. They do probably work on loading them last, but that's mainly because the overnight/2-day deadlines are 'hard' and (on the west coast) early. The FedEx jet is taking off from California at ~6:30pm... with OR without your package... so it makes sense to get THOSE packages out first. In contrast, nobody besides the driver really cares if a truck leaves at 9:30pm or 11:23pm on its cross-country trip.
They don't delay the non-prime to do the "paid" shipments first... the delay is because Amazon pre-purchases bulk FedEx shipping, and FedEx literally leaves trucks headed to various metro areas parked at Amazon's loading dock. Amazon's employees load up the trucks, and when a truck is full, Amazon calls them, FedEx brings a new truck, and drives away with the old truck that's now full and ready to head to its destination. If you live in a metro area that has lots of Amazon orders from that particular warehouse, the delay might be short. If you live in a metro area with few Amazon orders, it could take a lot longer. It all depends upon how long it takes for there to be enough orders to fill the truck.
> Don't forget unlimited $4 overnight. Do that once with a computer case and it pays for itself.
Or a generator. I have friends who literally bought a generator with prime shipping as a hurricane was making landfall, and had the generator on their front porch the following afternoon (obviously, they lived in an area where the power lines went down, but there was no major flooding or destruction).
The existence of Prime shipping is a game-changer to someone who lives in a place that gets hurricanes, because it enables you to buy things with overnight or 2-day shipping that would be cost-prohibitive to buy from them if you had to actually pay full price for that shipping. I don't even bother fighting the zoo at Home Depot, Target, or Sam's Club anymore before hurricanes... I just buy everything from Amazon, and come home from work to a pile of hurricane supplies waiting on my front doorstep the next day, a few hours before the hurricane makes landfall. Prime shipping is the whole reason why I now buy things like cat food, batteries, and lawnmower parts from Amazon, as opposed to books. Even places that offer free shipping can't compete, because THEY only offer free GROUND shipping, and if you want it upgraded to 2-day or next-day, you get hit with the full cost. With Amazon, 2-day is free, and overnight is only a few dollars more.
I just wish Amazon had a search option for "Only show me items that can be shipped in time to receive tomorrow". If there's any consistent logic to their cutoff times for same-day shipping, I have yet to figure out what it is. Sometimes it seems to be as late as 8:30pm, sometimes it seems to be 4pm, and it doesn't just seem to vary by warehouse... it literally seems to be a matter of blind, random good or bad luck that changes daily with no apparent rhyme or reason. Saturday and Sunday delivery are even more randomly variable. Few things are more frustrating than trying to do last-minute birthday present shopping on Friday as the deadlines are ticking away, and you can't even figure out what the deadlines ARE without clicking on the item and scrolling down. I swear to god, I'm going to write a program to do brute-force scraped-searches on Amazon called "OnlyTomorrow" that spoofs a browser, executes your search, then fetches the items and brute-force eliminates any that can't ship in time or that don't literally include the requested keywords in the title/description (my other pet peeve about Amazon... sometimes, its signal-to-noise ratio for searches is just horrendous, and the SNR seems to be the absolute WORST when you're panic-buying and desperately trying to beat the same-day shipping deadline).
I can't wait until somebody makes an amd64-architecture tablet with i7 performance. Why? Because THEN, it'll finally be possible to hack it into a thick-but-lightly-luggable laptop with a Model M keyboard (retrofitted with LiPo battery & bluetooth) hacked into an awesome case, and I won't be crippled into typing 50wpm the way I am with normal laptop keyboards (ok, maybe 65wpm on a Thinkpad w/classic keyboard), vs ~120 wpm with a Model M.
There might not be many of us, but mfrs. will have to pry our Model M keyboards from our cold, dead hands.They're like 1970s phone handsets -- they represent decades of industrial engineering & design to make a product that's *ideal*, as opposed to new keyboards whose only meaningful design feature is "lower manufacturing costs than last year's model"
> Seems sort of stupid to me to put generators in a basement
"Up north", basements are PRECISELY where you store things like gunky, messy generators (not necessarily with the full blessing of local officials) if you're in a big city like New York, in an old building that was built before elevators were mandatory and people still used coal for heat. They can't go on the roof, because they'd get damaged by the wind and rain. They can't go on the top floor, because it's the expensive penthouse. A newer building might have an integrated parking structure, in which case it would get tucked under a ramp or something... but a smaller building that doesn't occupy most of a block doesn't have that luxury.
My own dream peripheral for South Florida residents who live in tall buildings: a unit that combines a polyphase propane-fueled generator combined with 5,000 (give or take) BTU mini-split air conditioner. The outdoor part sits on the balcony, along with the 50-100 pound propane tank (both stored indoors during the storm itself). It contains the generator and air conditioner compressor. The indoor part is connected by a nonremovable 25-foot umbilical, and contains the controls, air conditioner blower, inverter/battery charger, power outlets, and connections for the lead-acid batteries. The umbilical contains the refrigerant flex pipes (well-insulated), wires for the polyphase A/C, and power wires for the compressor (it uses generator power to run, but takes advantage of battery power for extra reserve current when starting up).
It's propane, because you can safely store propane indoors & it doesn't go bad the way gasoline does. The exhaust is also cleaner and safer... an important point to consider in light of its proximity to living spaces. The outdoor part goes on the balcony (skyscraper) or on the ground outside a window, and the umbilical is oval-shaped so you can mostly close the sliding glass door or window and block the remaining gap with foam or a towel. The 25-foot umbilical length allows you to run it through the living room and enjoy A/C in an adjacent bedroom (it's quieter this way, too). It also allows you to put the outdoor unit on the ground, and put the indoor unit on a table just inside a second-story widow on a single-family home or townhouse. The generator runs when the air conditioner does, when there's a significant (> 250 watts) load, and/or the batteries are charging. On a cool night, the generator might shut down occasionally to conserve fuel. Whenever the load exceeds the generator's capacity, or slightly exceeds its 50% load capacity, it draws current from the batteries.
Imagine having a concrete-bunker warehouse full of them somewhere in western Broward County (near I-75/I-595/Sawgrass) and Orlando, and leasing them in a "fractional ownership" kind of way (with full service maintenance between storms) for $125/day (5-day minimum per year, 3-day minimum per storm, reduced to 1 day if you return it the next day unused so it can be put right back in the warehouse). For an optional nonrefundable $250 per year, you can have one set aside for you and guaranteed to be available, and get a discount of $25/day and have the minimum rental periods waived (since you're basically pre-paying 2 days and gambling on having at least one hurricane during the year). For another optional $100 per storm, you can get it delivered (voluntarily paying an additional $50 or $100 moves you to the front of the delivery queue, or lets you cut to the middle or front of the line if you go pick it up in person ).
Propane purchase is optional, but not included. This is a MAJOR potential profit, because bulk LPG is dirt cheap, and there are VERY FEW places in the urban parts of South Florida where you can fill your own tank at non-ripoff prices. The rest is all Blue Rhino LPG being sold in 2/3 full tanks at *criminal* prices, like $25 for a 20-pound tank with ~16 pounds of actual gas and proprietary valves that prevent you from ever refilling any tank you swap at a cheaper place going forward
No, you're the one who's omitting its most important point: if the manufacturer wants to deny a claim for some failure that would otherwise be covered by the warranty because you did ${something}, the burden is 100% on THEM to convince the FTC that the failure occurred BECAUSE you did ${something}, not merely that you DID ${something}.
Pre-MMWA, they just had to show the clause in the warranty that excluded coverage if you did ${something}, and prove that you DID ${something}. They could insert language invalidating the car's warranty if you installed an aftermarket stereo or changed the oil and used a non-approved aftermarket oil filter, even if the warranty claim were for something completely unrelated, like body rust.
If you root your phone or unlock its bootloader and reflash to a custom distro of Android, and your warranty claim for a delaminated OLED or cold solder joint on the USB port or headphone jack gets denied, the FTC will hang the vendor from the nearest tree, flog them until they're metaphorically nude and bloody, then hit them with a major fine for good measure after seeing to it that you've been adequately taken care of, because that's EXACTLY the kind of abuse MMWA is intended to stop.
If you brick your phone while trying to defeat the bootloader, MMWA doesn't officially apply. But you'll probably get your phone reflashed for free anyway, because JTAG reflashing is the first thing any repair center is going to try as their first diagnostic step. It comes down to risk vs reward for the manufacturer. If your phone was bricked, and JTAG reflashing fixed it, they'd have to be complete idiots to risk the FTC's wrath by trying to deny the warranty claim.
In theory, they'd have stronger case if you did something, like blow OTP fuses that rendered the phone permanently bricked, or caused real damage requiring real money to fix... but once again, consider that you're talking about a phone that represents (worst-case) a $200-300 hard loss for the manufacturer if they chuck it and send you a remanufactured replacement, vs thousands of dollars in potential fines and compliance costs if they choose to fight. In most cases, they'll probably hedge their bets and cut their losses by offering to replace the phone for $100-200 and call it a day. MMWA doesn't directly FORCE them do that, but it makes it too expensive for manufacturers to risk trying to profit from your phone's nonworking state.
MMWA doesn't give you the right to abuse a product and expect a free repair. It DOES, however, prohibit manufacturers from trying to use the warranty as a weapon to force compliance with its TOS or software licensing policies.
No, I have something better. A big honking 3600mAH Seidio extended battery w/NFC antenna and Active case with kickstand designed around the same battery. They cost a fortune, and they were worth every last penny, because I can now actively use my phone all day without ever having to settle for lag, a dim screen, or care about conserving battery power. A few times, I've gone to bed and forgot to plug it in (something that NEVER used to happen, because I was surgically-attached to my old phone's charger), and IT STILL WAS POWERED UP the next morning.
But since you ask, I also have a small pile of standard batteries I bought for something like $3 apiece from someone on eBay, just in case a hurricane knocks out the power for a few days and I need them for something. Well, also, because I didn't think the Seidio battery was going to be available until November, and I *really* needed them in the meantime. (I used to carry at least one spare with me at all times, and usually had a second one recharging from a charger from an inverter plugged into a cigarette lighter in my car as well).
The Nexus 4's lack of LTE is devastating. Its lack of microSD makes it unappealing. The nonremovable battery renders it completely moot.
It basically comes down to, "If you use your phone mainly for making voice calls, a wimpy Apple-like sealed battery is fine. If your phone is actively online from the moment you crawl out of bed in the morning, you hate lag, you dislike dim displays, and it's practically a second screen for your desktop PC or laptop, nothing less than 3000mAH will do,"
When our phones became sufficiently powerful, with sufficiently high-res screens, to just forget about 'mobile' sites and hit the 'desktop' site directly.;-)
Unless Congress has recently repealed the Magnuson-Moss Warranty Act, no. They can SAY your warranty is invalid, but they can also SAY the moon is made from green cheese. Under MMWA, unless they can prove that the failure was directly caused by unlocking the bootloader, their only recourse is to reflash your phone to stock with a JTAG (which they're going to do regardless) to get it back into a known state and verify that it's a hardware problem & not a software problem. MM does NOT require them to make the phone work using your custom ROM, nor do they have to return it to you after a warranty repair/replacement with anything besides the official firmware. It just means they can't refuse a warranty repair for a delaminated screen, or a volume button falling off, or a cracked USB/headphone jack, because your phone was rooted/reflashed.
Magnuson-Moss rocks. It's one of the most potent pro-consumer laws out there. Even APPLE, in their most arrogant AT&T-exclusive days, didn't dare to mess with the Federal Trade Commission. You don't even have to sue in court. You file a complaint with the FTC, and the manufacturer has a short period of time to respond explaining how they've remedied the situation, or explain why they believe the Magnuson Moss act doesn't apply. If they claim it doesn, and the FTC disagrees, they get fined a HUGE amount of money. Ergo, the manufacturer would have to be completely insane and suicidal to try and press a warranty denial unless their legal department assured them their defense was rock solid. Like I said, not even APPLE dares to fight with them. If you have someone in tech support tell you the warranty is void because you rooted or unlocked the bootloader, tell them you want to speak to their supervisor before filing a Magnuson Moss complaint. The moment a supervisor hears the phrase "Magnuson Moss", they're going to pull out the white gloves and offer you free blowjobs & chocolate to make you happy.
And before you ask... no, you can't be forced to waive your rights under it. If you buy a phone that has an American warranty, and they force you to waive it as a condition of getting an unlock code or something comparable, it's legally meaningless. Check 'yes', get the code, and submit your warranty claim anyway if you need to do it someday. By law, your rights under Magnuson Moss can NOT be waived (because if they could, everyone would just force you to waive them, and render it meaningless).
Bullshit. Just about every halfway-urban patch of Florida has LTE from AT&T and Verizon RIGHT NOW... and I'm not talking about a few neighborhoods in Miami, Orlando, and Tampa. AT&T has rock-solid LTE in Broward County, and I've had working AT&T LTE over the past month in Naples, Holmes Beach/Anna Maria Island (off the coast of Bradenton), and Ocoee (northwest Orlando, off 429, which is about as rural as you can get while still loosely describing the area as "Orlando"). I don't know about Verizon elsewhere in Florida, but I know their LTE is rock-solid in Miami and Fort Lauderdale from coworkers who use them.
> What has openness got to do with a micro sd slot!?
Internal SPI expansion bus that's trivially easy to program directly with minimal ceremony?
As an embedded hardware guy, I totally get warm fuzzies from SPI. It's just about the easiest low-ceremony bus on planet earth to use, and in a pinch you can even bitbang it with minimal effort. I know there's no room inside a microSD card for useful hardware thicker than a silicon wafer, but you could always use a fake microSD card connected to a ribbon cable to feed hardware built into a thicker replacement back.
>Better ad targeting or content matching algorithms definitely won't fix it
Maybe not, but you have to admit that if you're going to be force-fed ads, ads for computer hardware & home automation gear are several orders of magnitude less annoying than ads for feminine hygiene products, diapers, payday loans, personal injury lawyers, and [Romney|Obama].
That's a lame excuse. Put the LTE module on a daughtercard, sell the phone without it, and users with AT&T/Verizon/Sprint/MetroPCS can buy their carrier's LTE module, rip off the back, plug it in, and snap it back on. Or really, just put the RF amplifier and antenna on the daughtercard. I believe there's even a company in Japan that came up with a standard ~10 years ago for more or less this purpose (allowing users to buy a radio module from their carrier, and stick it into a phone made by someone else. If you think the US is bad, Japan is even worse when it comes to mobile networks with proprietary, closed phones).
There's no reason a phone with LTE can't be "open". The radio modem's firmware has nothing whatsoever to do with Android. It lives in its own firewalled universe. All Android does is pass the flash data to it when updating. If the radio modem wants to reject the new firmware, it can and will.
The whole issue with CDMA Nexi was a spat between Qualcomm (not wanting to release the source to peripherals on their SoC that had nothing to do with the actual phone/rf functionality), Verizon (not wanting open devices, period), and the reality that the existence of semi-open firmware for Sprint phones makes hacking Verizon phones with more or less identical hardware a lot easier. Verizon was vehemently opposed, Sprint was either indifferent (or too broke and poor to matter), and Qualcomm just stonewalled Google until they got frustrated and gave up.
This application wants to send a text message (_content_) to _number_. How should I handle it?
[ ] Deny, and ask me for permission next time. [ ] Deny, and disable this app ([ ]until (*)tomorrow, ( )next week, ( )next month) [ ] Deny, and uninstall this app [ ] Allow this time [ ] Allow until tomorrow [ ] Allow until next week [ ] Allow forever [ ] Allow if (_click to select app_), which implements IGatekeeper, says it's OK to allow.
The gaping hole in Android's security model is the fact that in order to have an app that fetches location-based ads over the internet, uses wi-fi (instead of GPS) for coarse location, and has the ability to pause when the phone rings or cooperate nicely with alerts and other apps, you basically have to give the app the right to do almost everything up to and including scrape your phone logs and dump them over the internet to the developer's server, then eavesdrop on your LAN's traffic and report it as well.
I don't have time to repost the whole essay I've written a few times detailing a provider-agnostic framework for adserving that keeps apps from leaking private user info by moving responsibility by proxying the network calls to fetch the ad content through Android itself in a way that allows users to say, "I trust Android to not leak my private info, but not this specific app... I'll allow this app to treat Android's new adserver API like a semi-black box to fetch ads in a way that prevents apps from injecting values not carved into stone in android-manifest.xml, and has Android itself inject sensitive values so the app itself can't touch them, and makes the requests in non-realtime with somewhat randomized timing through Google's adproxy (or a trusted CDN, for larger advertising agencies with the resources to pay someone like Akamai) that masquerades the user's IP address (so developers can't comb through logs and match up ad requests with IP addresses).
If Android did something like this, the laundry list of permissions that 98% of modern Android apps end up requiring could be concisely boiled down to:
* Display anonymized, location-based ads fetched over the internet in a way that does not reveal your current IP address or personally-identifiable information to advertisers or the app's developer, and does not allow apps to use it as a back door to leak information over the internet by injecting runtime values specified by the application itself into the ad request or by varying the timing of its network requests to convey private information to a remote server.
The requirement that values either be filled in by Android itself (in a black-box manner that keeps the values away from the app) or declared immutably via android-manifest.xml, and slightly-randomized non-realtime ad-fetching and timing is necessary to keep apps from using runtime values or timing attacks to leak information. If ads are fetched on demand by the app, the developer could modulate the request timing itself to convey one bit of data at a time, over a long period (ex: requesting new ad within 1 minute of last request == 1, requesting new ad after 2 minutes of last request == 0, requesting new ad after 3+ minutes = escape, resume as directed by the next few bits... over the span of an hour, you could leak 30-60 bits of data).
The hard part would be making it vendor-agnostic and not handing an ad monopoly to Google, without excluding ad agencies who don't have the resources of Akamai (hence, the transparent trusted anonymizing proxy for fetching the ad data itself).
Exactly. Imagine the mess South Florida would have if a freak winter storm dumped 2 inches of real mid-afternoon snow on downtown Miami and the surrounding 3 counties, and it kept coming down all night so that we woke up the next morning to a city where every road was impassable to anything less than a SUV or truck, seriously dangerous regardless, and every vehicle that was outside overnight had ice crusted over the windshield wipers. We'd have people getting electrocuted trying to melt ice on the windshield with blow dryers (until the weight of the ice caused the power lines to fall down), and I shudder to imagine the carnage on I-95 and 836 when drivers who can't even avoid accidents during afternoon rainstorms suddenly had to deal with ICE.
A category 1 hurricane making landfall in South Florida is like a "Snow Day" in Cleveland or Buffalo -- work from home today, limp and tipetoe around tomorrow, life as normal on day 3. An afternoon snowstorm that persists into the night would shut down South Florida for almost a week, and probably cause more deaths than a landfalling hurricane.
Errr... you have a point insofar as people more or less ignoring NOAA's own weather reports in preference to commercial alternatives goes, but you're dead wrong about satellites and radar. Those commercial stations depend upon the constellation of satellites and array of radar sites that are operated by NOAA, regardless of whether the actual construction was done by a government employee or contracted out to someone.
Yes, there are a few TV stations that have their own X-band weather radar, but they're mostly eye candy for the local audience (limited range, limited features). For their real forecasting work, they grab the level 2 data from NOAA's radar. You can argue about the need for NOAA's meteorologists, but you'd have to be completely delusional to think their data reconnaissance services could be adequately replaced by commercial alternatives. Not even The Weather Channel has the resources to send out the hurricane hunter planes, let alone maintain its own constellation of satellites and operate an array of radar sites with anything close to the scope and capabilities of NEXRAD and TDWR.
Would a 100% commercial enterprise be more efficient? Certainly. However, efficiency isn't everything. Availability matters, too. It's "efficient" (for the profits of an investor-owned power company) to just accept rolling blackouts for a day or two per year, instead of "over-building" their capacity to make sure it never, ever happens. That doesn't change the fact that it really sucks to be in one of those areas when the blackouts happen, and the cost of coming up with your own backup power is several orders of magnitude more than what it would have cost in higher monthly bills had the network just been engineered to a higher standard in the first place. A private company would roll the dice and risk the loss of a satellite or two for a few years, even though the marginal cost of the spares (spread across ~300 million taxpayers) is next to nothing. A private company can't do that, because it only has a few (compared to 300 million) paying customers, so the extra satellite goes from an extra cent or two per year in taxes to doubled (or more) monthly/annual subscription fees.
Plus, I should probably point out that weather affects interstate commerce in a major way, and is probably one of the most constitutionally-unambiguous responsibilities OF the federal government to handle.
In any case, if you think a spare satellite or two is expensive, take a wild guess how much it would cost to relocate SBX-1 to some location in the middle of the Gulf of Mexico, or 200 miles off the east coast between Jacksonville and Cape Hatteras, to do stand-in duty as the most expensive weather radar platform in the history of meteorology (not to mention the damage it would almost certainly sustain if it were literally put in the path of a half-dozen hurricanes in a single year).
The single hardest thing I've had to do in my house is run wires horizontally. Pulling wires vertically is relatively easy, but fishing them horizontally is a nightmare. Run horizontal conduit across every room, and terminate it at low-voltage frames that face opposite sides of the wall between alternating studs (to avoid having blank wallplates every 2 feet in any one room. If the room backs up against a closet, put frames in the room where you want them, and put the rest of them on the 'closet' side of the wall. That way, if you need to pull wires horizontally in the future, you can just remove the faceplates and pull the cat5/6/fiber/whatever across the room. On each wall, have a junction leading to another conduit that runs to a central wiring area. It's OK and appropriate to have two or more such areas... say, one near the room where your home theater system is located (so you can put most of your gear there, so you can have direct physical access to it from the location where most of your real-world interactive viewing will occur, and feed other rooms semi-directly from there), and at least one on each floor.
I can't emphasize the "have multiple wiring paths and crossover points" enough. I've seem some people go to patently absurd lengths so they could get ideological about having "home runs" for everything... then complain that their system is too brittle, or they had to buy a thousand-dollar HDMI extender instead of a $60 one, because they insisted on sending the signal on a 700-foot detour to some central location instead of just pulling a wire through the wall and feeding the jack on the other side. There's a time and place for both. For networks, go ahead and homerun. For high-bandwidth stuff that has serious, expensive-to-fix problems as the distance increases beyond a hundred feet or so, run point to point. Build your conduit so you can do both. Just because you CAN solve certain problems by throwing expensive hardware at them doesn't mean it's necessarily the best idea if you aren't faced with insurmountable physical constraints. Wires are cheap. Managed VLAN-capable gigabit switches with IGMP 3 snooping aren't.
Ask yourself: if, 7 years from now, you somehow had to get a wire from (some random point in a room) to (some random point in another room on the other side of the house), how would you do it? Could you do it without major surgery as, at worst, a weekend project that might involve cutting a new hole or two for the faceplate frame and some vertical fishing? Because if you can achieve that with conduit, you'll be set for anything you can think of for the next 40 years. Don't even TRY to anticipate the kinds of cables you'll need to run. You'll be wrong, and end up pulling a shit-ton of cable that will end up being sub-optimal for whatever you had in mind, anyway. Run horizontal conduit, interconnect walls with junction points around the house, and sleep well knowing you're ready for anything. I know firsthand -- I spent a fortune, and weeks, pulling shielded cat5e cable after work so I'd be able to run component video over cat5e without interference. I ended up buying $50 HDMI-over-1-cat5e extenders, and all the shielding (god, it was a pain to terminate) ended up being moot.
Remember, high- and low-voltage can't share conduit or boxes (unless divided). I personally recommend running 4 conduits across each wall... 2 for LV, 2 for HV (possibily substituting Romex for one of the HV conduits for the initial build). Use one conduit in each pair for your "phase 1" wiring connecting your "phase 1" boxes, and use the second conduit in each pair to run future wires from "phase 1" boxes to other locations where you can't bring yourself to put blank wallplates right now. By using the second conduit for "the last 2/4/6 feet", you'll minimize the number of wires that have to be pulled out and re-pulled if/when future cuts to condudit #2 are necessary.
For high-voltage wiring where you're putting new boxes, put the deepest boxes you can, with box extenders if possible. If you're putting
Sigh. It's things like that that make me start to seriously think that Google is losing its way, kind of like Microsoft did between NT and Win2k... the point when the original innovators are now too rich to actually be useful as employees anymore, but are politically untouchable & starting to drag everything else down by virtue of their own de-facto disinterest in their official job... delegating everything they can to people whom they refuse to actually give real decision-making authority, then scrambling to cobble things together at the last minute when push comes to shove before dropping off the earth for days/weeks at a time to go sailing, climb Mt. Everest, visit Antarctica, and travel to Baikonur for a few weeks to see whether they have what it takes (besides a few million dollars) to become a Commercial Cosmonaut.
The Nexus One was groundbreaking. The Nexus S was thoroughly 'meh' (admittedly, compounded by the fact that Sprint took so long to get it, by the time it was available, it literally WAS last year's phone), and the Galaxy Nexus was quickly overshadowed by the Galaxy S3, and equally 'meh' by the time Verizon's exclusivity ended.
Arguably, the fact that Google even ALLOWED Verizon to have exclusivity to both the name and franchise, despite having the gall to lock the phone's bootloader and take a corporate dump on everything it means for a device to BE a "Nexus", is more evidence that someone in their Android group has either completely lost touch with reality, or just plain didn't care because he was too busy training for space camp.
I've said it a thousand times, but what Google really needs to do is just pick various manufacturers' best of breed phones for the "Nexus Treatment", and make it so you'd just go and buy a phone like the Galaxy S3, OneX, or RazrMax, then download Google's official AOSP build for it, plug in the USB cable and connect it to your computer, then blow away the carrier's stock ROM with the official Google-blessed AOSP Nexus kernel and distro for it, instead (kind of like buying a laptop with Windows Starter Edition, then blowing it away with Ubuntu or Mint). Otherwise, Google's going to end up having their whole Nexus platform end up turning into "last year's hardware, slightly reheated and limping along with next year's software".
I doubt it, or at least doubt that they'd instantly go into such a mode the moment commercial power were lost. Otherwise, they'd have lots of REALLY angry customers who'd be furious about being unable to call anyone besides 911.
At worst, I could see them load-shedding the fiber or microwave link and falling back to a single T1 if the power situation became dire (or the fiber/microwave link lost its connectivity due to some upstream problem anyway), but prohibiting everything besides 911 would cause them more problems than it would solve. Especially when you consider that during a real hurricane, calling 911 is almost pointless anyway, because they aren't going to risk the lives of police/paramedics/firefighters by sending them out in 100+mph winds with trees and live power lines down on the road.
Still, I guess this means I should also look into buying an old rootable & reflashable GSM phone on eBay while I'm at it. I remember reading somewhere that there was at least one guy working on implementing a v.92 "VOIP-mode" 2400/4800/9600 baud modem for Android that would work by taking direct control over the phone's audio hardware and bitbanging it in realtime, just like a Winmodem would. I think the original intent was to let you do things like send faxes to real-world fax machines straight from your phone, but I don't see why it couldn't be used to connect to a modem bank somewhere, too.
Before someone mentions that modems don't work over cell phones, I should mention that there ARE a couple of modes that were newly defined sometime around v.92 or v94 that are basically high-bandwidth low-bitrate modes that try to trick VoIP or wireless codecs into thinking that they're speech plus loud background noise that can't be separated out or compressed away, and overwhelm the codec into giving up and encoding the content at a higher bitrate than it would ever use for straight 2400 baud.
The idea is that if you try to send straight 2400 baud tones through something like a GSM codec, it'll try to encode them with the equivalent of 1200 bits per second and you'll end up with garbage. But, if you encode 2,400 bits per second along with 48,000 bits per second of Reed-Solomon error correction, and do it in a way that totally overwhelms the compression algorithm, it'll grudgingly end up encoding it with a much higher bitrate, and leave you with enough to recover the original 2400 baud data at the other end. It's mostly used now for fax machines and credit card processing, but in theory it could also be used for straight dialup internet.
What's your exact model number? I'm intrigued, and might want to buy one. I was under the impression that TrippLite's UPSes were among the most intolerant (of frequency deviation) out there.
The most tolerant UPSes I'm aware of are supposedly APC's, IF you explicitly kick them into 'tolerant mode'. I believe they're still rigid about frequency, but will tolerate voltage sags a little longer before kicking in, so it might prevent thrashing with an inverter-type generator.
Side note: only APC's high-end and business-class UPSes officially support 'tolerant mode', but from what I've seen, even their consumer-grade models have the capability... it's just not something you can enable from their official management/monitoring software. However, if you open a serial port and blindly send the right sequence of bytes to the UPS, you can probably enable it on their consumer models, too. I know that I have an ES350, and was able to trigger it just fine. Unfortunately, it didn't fix the problem for me (it made it thrash a tiny bit less, but it still thrashed).
Slashdot Mirror
Are you serious? I've accidentally "liked", flagged, thumbed-up/down, and other actions while browsing sites with my Android phones over the years, I think developers should be taken out and SHOT for allowing actions with visible, public, and lasting consequences to EVER get triggered by a single touch. They should either require separate confirmation, or give you at least a few seconds to click "ohshit, no, I didn't mean that!!!" to retract/undo them before anyone ever finds out.
PhpBB and other web forums, along with Facebook, are the absolute WORST in this regard. They send out email notifications instantly, so even if you try to retract it, they're going to know you did it anyway.
It should be absolutely, unambiguously legal to photograph a ballot that's in your possession, and publish it however you see fit without restriction.
HOWEVER... any ballot that gets photographed should be deemed "tainted", destroyed (or unambiguously rendered uncountable), and replaced with a fresh ballot (possibly after informing the errant voter that photographed ballots may not be cast, nor may ballots with identifying marks that could be used to identify the ballot as "theirs" in a recount(*). Otherwise, friends, family members, unions, employers, or anyone else could try and coerce members into voting "correctly" by pressuring them to prove that they voted the "right way".
There's a fine line when it comes to appropriate enforcement. It would be wrong to punish somebody who tries to trick officials into accepting his or her photographed ballot... you can only do so much to save someone from themselves... but it WOULD be appropriate to treat it like a "CAPTCHA" situation to prevent someone from attempting to launch a denial of service attack:
The first time someone is caught photographing his or her ballot, they're informed that they have the right to photograph their ballot, but the ballot itself is now legally tainted and can not be cast. It must be destroyed, and they can vote again from a fresh ballot. They'll be informed that they can photograph at their leisure, but would be asked to show consideration for others waiting in line and suggested that they continue their photography at a location where it won't interfere with other voters waiting in line, like the table set up "over there, in the corner".
If they do it a secnod time, they'll be required to watch a short (5-10 minute) video about the importance of secret ballots in a democracy, and asked to place all photography devices in a tamper-evident bag prior to receiving their new ballot. They're told that if the bag is tampered with, or they're observed attempting to photograph their ballot with yet another device, they'll be subjected to additional requirements before they'll be allowed to make their third attempt and beyond.
If they do it a third time, they'll be required to watch the video, wait in line, watch the video again, seal their photography devices in a tamper-evident bag, and write a paragraph expressing their understanding of the rules and their reason. Then vote again on a clean ballot.
If they get caught again, on attempts 4 and beyond, they'll have to watch the video and wait in line once time for each violation beyond the third. So, attempt #4 means watch, wait, watch, wait, watch, write, and vote. Attempt #5 means watch, wait, watch, wait, watch, wait, watch, write, vote. And so on. If they're on attempt #3 or beyond when the polls officially close and the ballot boxes are sealed, and they're stuck waiting, they're out of luck. Essentially, everyone gets one free attempt to photograph and do-over without risk of punishment (beyond having to fill out the ballot again), and you have to actively screw up one more time beyond that before you might end up not being able to vote at all. A hard limit is a serious matter, but if you have someone who's actively gone through the cycle more than twice, it's safe to assume at that point that they're so completely fucked in the head (or coerced, or hellbent on making a political statement by the act itself, or all the above) that it's just a waste of everyone's time to accommodate them further at that point.
This is something that's only going to become a bigger and bigger problem over time. It's also the #1 reason why we can never have internet voting -- there's no way to guarantee that somebody isn't standing over the voter's shoulder, ready to reward them for voting "properly", or punish (real, or implied) them if they don't cooperate and vote the right way.
(*) I'd go a step further, and require that ballots be designed in a way that isolates them into clusters of comparable races, with a maximum of 3 or 4 races per physical ballot-sheet being r
The overnight & 2-day boxes don't go anywhere *near* the 'economy' trucks. Think of the economy trucks as a bulk shipping crate with wheels, whose workflow is completely independent of faster methods. They do probably work on loading them last, but that's mainly because the overnight/2-day deadlines are 'hard' and (on the west coast) early. The FedEx jet is taking off from California at ~6:30pm... with OR without your package... so it makes sense to get THOSE packages out first. In contrast, nobody besides the driver really cares if a truck leaves at 9:30pm or 11:23pm on its cross-country trip.
They don't delay the non-prime to do the "paid" shipments first... the delay is because Amazon pre-purchases bulk FedEx shipping, and FedEx literally leaves trucks headed to various metro areas parked at Amazon's loading dock. Amazon's employees load up the trucks, and when a truck is full, Amazon calls them, FedEx brings a new truck, and drives away with the old truck that's now full and ready to head to its destination. If you live in a metro area that has lots of Amazon orders from that particular warehouse, the delay might be short. If you live in a metro area with few Amazon orders, it could take a lot longer. It all depends upon how long it takes for there to be enough orders to fill the truck.
> Don't forget unlimited $4 overnight. Do that once with a computer case and it pays for itself.
Or a generator. I have friends who literally bought a generator with prime shipping as a hurricane was making landfall, and had the generator on their front porch the following afternoon (obviously, they lived in an area where the power lines went down, but there was no major flooding or destruction).
The existence of Prime shipping is a game-changer to someone who lives in a place that gets hurricanes, because it enables you to buy things with overnight or 2-day shipping that would be cost-prohibitive to buy from them if you had to actually pay full price for that shipping. I don't even bother fighting the zoo at Home Depot, Target, or Sam's Club anymore before hurricanes... I just buy everything from Amazon, and come home from work to a pile of hurricane supplies waiting on my front doorstep the next day, a few hours before the hurricane makes landfall. Prime shipping is the whole reason why I now buy things like cat food, batteries, and lawnmower parts from Amazon, as opposed to books. Even places that offer free shipping can't compete, because THEY only offer free GROUND shipping, and if you want it upgraded to 2-day or next-day, you get hit with the full cost. With Amazon, 2-day is free, and overnight is only a few dollars more.
I just wish Amazon had a search option for "Only show me items that can be shipped in time to receive tomorrow". If there's any consistent logic to their cutoff times for same-day shipping, I have yet to figure out what it is. Sometimes it seems to be as late as 8:30pm, sometimes it seems to be 4pm, and it doesn't just seem to vary by warehouse... it literally seems to be a matter of blind, random good or bad luck that changes daily with no apparent rhyme or reason. Saturday and Sunday delivery are even more randomly variable. Few things are more frustrating than trying to do last-minute birthday present shopping on Friday as the deadlines are ticking away, and you can't even figure out what the deadlines ARE without clicking on the item and scrolling down. I swear to god, I'm going to write a program to do brute-force scraped-searches on Amazon called "OnlyTomorrow" that spoofs a browser, executes your search, then fetches the items and brute-force eliminates any that can't ship in time or that don't literally include the requested keywords in the title/description (my other pet peeve about Amazon... sometimes, its signal-to-noise ratio for searches is just horrendous, and the SNR seems to be the absolute WORST when you're panic-buying and desperately trying to beat the same-day shipping deadline).
I can't wait until somebody makes an amd64-architecture tablet with i7 performance. Why? Because THEN, it'll finally be possible to hack it into a thick-but-lightly-luggable laptop with a Model M keyboard (retrofitted with LiPo battery & bluetooth) hacked into an awesome case, and I won't be crippled into typing 50wpm the way I am with normal laptop keyboards (ok, maybe 65wpm on a Thinkpad w/classic keyboard), vs ~120 wpm with a Model M.
There might not be many of us, but mfrs. will have to pry our Model M keyboards from our cold, dead hands.They're like 1970s phone handsets -- they represent decades of industrial engineering & design to make a product that's *ideal*, as opposed to new keyboards whose only meaningful design feature is "lower manufacturing costs than last year's model"
> Seems sort of stupid to me to put generators in a basement
"Up north", basements are PRECISELY where you store things like gunky, messy generators (not necessarily with the full blessing of local officials) if you're in a big city like New York, in an old building that was built before elevators were mandatory and people still used coal for heat. They can't go on the roof, because they'd get damaged by the wind and rain. They can't go on the top floor, because it's the expensive penthouse. A newer building might have an integrated parking structure, in which case it would get tucked under a ramp or something... but a smaller building that doesn't occupy most of a block doesn't have that luxury.
My own dream peripheral for South Florida residents who live in tall buildings: a unit that combines a polyphase propane-fueled generator combined with 5,000 (give or take) BTU mini-split air conditioner. The outdoor part sits on the balcony, along with the 50-100 pound propane tank (both stored indoors during the storm itself). It contains the generator and air conditioner compressor. The indoor part is connected by a nonremovable 25-foot umbilical, and contains the controls, air conditioner blower, inverter/battery charger, power outlets, and connections for the lead-acid batteries. The umbilical contains the refrigerant flex pipes (well-insulated), wires for the polyphase A/C, and power wires for the compressor (it uses generator power to run, but takes advantage of battery power for extra reserve current when starting up).
It's propane, because you can safely store propane indoors & it doesn't go bad the way gasoline does. The exhaust is also cleaner and safer... an important point to consider in light of its proximity to living spaces. The outdoor part goes on the balcony (skyscraper) or on the ground outside a window, and the umbilical is oval-shaped so you can mostly close the sliding glass door or window and block the remaining gap with foam or a towel. The 25-foot umbilical length allows you to run it through the living room and enjoy A/C in an adjacent bedroom (it's quieter this way, too). It also allows you to put the outdoor unit on the ground, and put the indoor unit on a table just inside a second-story widow on a single-family home or townhouse. The generator runs when the air conditioner does, when there's a significant (> 250 watts) load, and/or the batteries are charging. On a cool night, the generator might shut down occasionally to conserve fuel. Whenever the load exceeds the generator's capacity, or slightly exceeds its 50% load capacity, it draws current from the batteries.
Imagine having a concrete-bunker warehouse full of them somewhere in western Broward County (near I-75/I-595/Sawgrass) and Orlando, and leasing them in a "fractional ownership" kind of way (with full service maintenance between storms) for $125/day (5-day minimum per year, 3-day minimum per storm, reduced to 1 day if you return it the next day unused so it can be put right back in the warehouse). For an optional nonrefundable $250 per year, you can have one set aside for you and guaranteed to be available, and get a discount of $25/day and have the minimum rental periods waived (since you're basically pre-paying 2 days and gambling on having at least one hurricane during the year). For another optional $100 per storm, you can get it delivered (voluntarily paying an additional $50 or $100 moves you to the front of the delivery queue, or lets you cut to the middle or front of the line if you go pick it up in person ).
Propane purchase is optional, but not included. This is a MAJOR potential profit, because bulk LPG is dirt cheap, and there are VERY FEW places in the urban parts of South Florida where you can fill your own tank at non-ripoff prices. The rest is all Blue Rhino LPG being sold in 2/3 full tanks at *criminal* prices, like $25 for a 20-pound tank with ~16 pounds of actual gas and proprietary valves that prevent you from ever refilling any tank you swap at a cheaper place going forward
No, you're the one who's omitting its most important point: if the manufacturer wants to deny a claim for some failure that would otherwise be covered by the warranty because you did ${something}, the burden is 100% on THEM to convince the FTC that the failure occurred BECAUSE you did ${something}, not merely that you DID ${something}.
Pre-MMWA, they just had to show the clause in the warranty that excluded coverage if you did ${something}, and prove that you DID ${something}. They could insert language invalidating the car's warranty if you installed an aftermarket stereo or changed the oil and used a non-approved aftermarket oil filter, even if the warranty claim were for something completely unrelated, like body rust.
If you root your phone or unlock its bootloader and reflash to a custom distro of Android, and your warranty claim for a delaminated OLED or cold solder joint on the USB port or headphone jack gets denied, the FTC will hang the vendor from the nearest tree, flog them until they're metaphorically nude and bloody, then hit them with a major fine for good measure after seeing to it that you've been adequately taken care of, because that's EXACTLY the kind of abuse MMWA is intended to stop.
If you brick your phone while trying to defeat the bootloader, MMWA doesn't officially apply. But you'll probably get your phone reflashed for free anyway, because JTAG reflashing is the first thing any repair center is going to try as their first diagnostic step. It comes down to risk vs reward for the manufacturer. If your phone was bricked, and JTAG reflashing fixed it, they'd have to be complete idiots to risk the FTC's wrath by trying to deny the warranty claim.
In theory, they'd have stronger case if you did something, like blow OTP fuses that rendered the phone permanently bricked, or caused real damage requiring real money to fix... but once again, consider that you're talking about a phone that represents (worst-case) a $200-300 hard loss for the manufacturer if they chuck it and send you a remanufactured replacement, vs thousands of dollars in potential fines and compliance costs if they choose to fight. In most cases, they'll probably hedge their bets and cut their losses by offering to replace the phone for $100-200 and call it a day. MMWA doesn't directly FORCE them do that, but it makes it too expensive for manufacturers to risk trying to profit from your phone's nonworking state.
MMWA doesn't give you the right to abuse a product and expect a free repair. It DOES, however, prohibit manufacturers from trying to use the warranty as a weapon to force compliance with its TOS or software licensing policies.
> Do you actually carry multiple batteries?
No, I have something better. A big honking 3600mAH Seidio extended battery w/NFC antenna and Active case with kickstand designed around the same battery. They cost a fortune, and they were worth every last penny, because I can now actively use my phone all day without ever having to settle for lag, a dim screen, or care about conserving battery power. A few times, I've gone to bed and forgot to plug it in (something that NEVER used to happen, because I was surgically-attached to my old phone's charger), and IT STILL WAS POWERED UP the next morning.
But since you ask, I also have a small pile of standard batteries I bought for something like $3 apiece from someone on eBay, just in case a hurricane knocks out the power for a few days and I need them for something. Well, also, because I didn't think the Seidio battery was going to be available until November, and I *really* needed them in the meantime. (I used to carry at least one spare with me at all times, and usually had a second one recharging from a charger from an inverter plugged into a cigarette lighter in my car as well).
The Nexus 4's lack of LTE is devastating. Its lack of microSD makes it unappealing. The nonremovable battery renders it completely moot.
It basically comes down to, "If you use your phone mainly for making voice calls, a wimpy Apple-like sealed battery is fine. If your phone is actively online from the moment you crawl out of bed in the morning, you hate lag, you dislike dim displays, and it's practically a second screen for your desktop PC or laptop, nothing less than 3000mAH will do,"
When our phones became sufficiently powerful, with sufficiently high-res screens, to just forget about 'mobile' sites and hit the 'desktop' site directly. ;-)
Unless Congress has recently repealed the Magnuson-Moss Warranty Act, no. They can SAY your warranty is invalid, but they can also SAY the moon is made from green cheese. Under MMWA, unless they can prove that the failure was directly caused by unlocking the bootloader, their only recourse is to reflash your phone to stock with a JTAG (which they're going to do regardless) to get it back into a known state and verify that it's a hardware problem & not a software problem. MM does NOT require them to make the phone work using your custom ROM, nor do they have to return it to you after a warranty repair/replacement with anything besides the official firmware. It just means they can't refuse a warranty repair for a delaminated screen, or a volume button falling off, or a cracked USB/headphone jack, because your phone was rooted/reflashed.
Magnuson-Moss rocks. It's one of the most potent pro-consumer laws out there. Even APPLE, in their most arrogant AT&T-exclusive days, didn't dare to mess with the Federal Trade Commission. You don't even have to sue in court. You file a complaint with the FTC, and the manufacturer has a short period of time to respond explaining how they've remedied the situation, or explain why they believe the Magnuson Moss act doesn't apply. If they claim it doesn, and the FTC disagrees, they get fined a HUGE amount of money. Ergo, the manufacturer would have to be completely insane and suicidal to try and press a warranty denial unless their legal department assured them their defense was rock solid. Like I said, not even APPLE dares to fight with them. If you have someone in tech support tell you the warranty is void because you rooted or unlocked the bootloader, tell them you want to speak to their supervisor before filing a Magnuson Moss complaint. The moment a supervisor hears the phrase "Magnuson Moss", they're going to pull out the white gloves and offer you free blowjobs & chocolate to make you happy.
And before you ask... no, you can't be forced to waive your rights under it. If you buy a phone that has an American warranty, and they force you to waive it as a condition of getting an unlock code or something comparable, it's legally meaningless. Check 'yes', get the code, and submit your warranty claim anyway if you need to do it someday. By law, your rights under Magnuson Moss can NOT be waived (because if they could, everyone would just force you to waive them, and render it meaningless).
> LTE is not ready yet
Bullshit. Just about every halfway-urban patch of Florida has LTE from AT&T and Verizon RIGHT NOW... and I'm not talking about a few neighborhoods in Miami, Orlando, and Tampa. AT&T has rock-solid LTE in Broward County, and I've had working AT&T LTE over the past month in Naples, Holmes Beach/Anna Maria Island (off the coast of Bradenton), and Ocoee (northwest Orlando, off 429, which is about as rural as you can get while still loosely describing the area as "Orlando"). I don't know about Verizon elsewhere in Florida, but I know their LTE is rock-solid in Miami and Fort Lauderdale from coworkers who use them.
If it fits in the rear pocket of an adult male while encased in an Otterbox Defender, it's not too big :-)
> What has openness got to do with a micro sd slot!?
Internal SPI expansion bus that's trivially easy to program directly with minimal ceremony?
As an embedded hardware guy, I totally get warm fuzzies from SPI. It's just about the easiest low-ceremony bus on planet earth to use, and in a pinch you can even bitbang it with minimal effort. I know there's no room inside a microSD card for useful hardware thicker than a silicon wafer, but you could always use a fake microSD card connected to a ribbon cable to feed hardware built into a thicker replacement back.
>Better ad targeting or content matching algorithms definitely won't fix it
Maybe not, but you have to admit that if you're going to be force-fed ads, ads for computer hardware & home automation gear are several orders of magnitude less annoying than ads for feminine hygiene products, diapers, payday loans, personal injury lawyers, and [Romney|Obama].
That's a lame excuse. Put the LTE module on a daughtercard, sell the phone without it, and users with AT&T/Verizon/Sprint/MetroPCS can buy their carrier's LTE module, rip off the back, plug it in, and snap it back on. Or really, just put the RF amplifier and antenna on the daughtercard. I believe there's even a company in Japan that came up with a standard ~10 years ago for more or less this purpose (allowing users to buy a radio module from their carrier, and stick it into a phone made by someone else. If you think the US is bad, Japan is even worse when it comes to mobile networks with proprietary, closed phones).
There's no reason a phone with LTE can't be "open". The radio modem's firmware has nothing whatsoever to do with Android. It lives in its own firewalled universe. All Android does is pass the flash data to it when updating. If the radio modem wants to reject the new firmware, it can and will.
The whole issue with CDMA Nexi was a spat between Qualcomm (not wanting to release the source to peripherals on their SoC that had nothing to do with the actual phone/rf functionality), Verizon (not wanting open devices, period), and the reality that the existence of semi-open firmware for Sprint phones makes hacking Verizon phones with more or less identical hardware a lot easier. Verizon was vehemently opposed, Sprint was either indifferent (or too broke and poor to matter), and Qualcomm just stonewalled Google until they got frustrated and gave up.
This application wants to send a text message (_content_) to _number_. How should I handle it?
[ ] Deny, and ask me for permission next time.
[ ] Deny, and disable this app ([ ]until (*)tomorrow, ( )next week, ( )next month)
[ ] Deny, and uninstall this app
[ ] Allow this time
[ ] Allow until tomorrow
[ ] Allow until next week
[ ] Allow forever
[ ] Allow if (_click to select app_), which implements IGatekeeper, says it's OK to allow.
The gaping hole in Android's security model is the fact that in order to have an app that fetches location-based ads over the internet, uses wi-fi (instead of GPS) for coarse location, and has the ability to pause when the phone rings or cooperate nicely with alerts and other apps, you basically have to give the app the right to do almost everything up to and including scrape your phone logs and dump them over the internet to the developer's server, then eavesdrop on your LAN's traffic and report it as well.
I don't have time to repost the whole essay I've written a few times detailing a provider-agnostic framework for adserving that keeps apps from leaking private user info by moving responsibility by proxying the network calls to fetch the ad content through Android itself in a way that allows users to say, "I trust Android to not leak my private info, but not this specific app... I'll allow this app to treat Android's new adserver API like a semi-black box to fetch ads in a way that prevents apps from injecting values not carved into stone in android-manifest.xml, and has Android itself inject sensitive values so the app itself can't touch them, and makes the requests in non-realtime with somewhat randomized timing through Google's adproxy (or a trusted CDN, for larger advertising agencies with the resources to pay someone like Akamai) that masquerades the user's IP address (so developers can't comb through logs and match up ad requests with IP addresses).
If Android did something like this, the laundry list of permissions that 98% of modern Android apps end up requiring could be concisely boiled down to:
* Display anonymized, location-based ads fetched over the internet in a way that does not reveal your current IP address or personally-identifiable information to advertisers or the app's developer, and does not allow apps to use it as a back door to leak information over the internet by injecting runtime values specified by the application itself into the ad request or by varying the timing of its network requests to convey private information to a remote server.
The requirement that values either be filled in by Android itself (in a black-box manner that keeps the values away from the app) or declared immutably via android-manifest.xml, and slightly-randomized non-realtime ad-fetching and timing is necessary to keep apps from using runtime values or timing attacks to leak information. If ads are fetched on demand by the app, the developer could modulate the request timing itself to convey one bit of data at a time, over a long period (ex: requesting new ad within 1 minute of last request == 1, requesting new ad after 2 minutes of last request == 0, requesting new ad after 3+ minutes = escape, resume as directed by the next few bits... over the span of an hour, you could leak 30-60 bits of data).
The hard part would be making it vendor-agnostic and not handing an ad monopoly to Google, without excluding ad agencies who don't have the resources of Akamai (hence, the transparent trusted anonymizing proxy for fetching the ad data itself).
Exactly. Imagine the mess South Florida would have if a freak winter storm dumped 2 inches of real mid-afternoon snow on downtown Miami and the surrounding 3 counties, and it kept coming down all night so that we woke up the next morning to a city where every road was impassable to anything less than a SUV or truck, seriously dangerous regardless, and every vehicle that was outside overnight had ice crusted over the windshield wipers. We'd have people getting electrocuted trying to melt ice on the windshield with blow dryers (until the weight of the ice caused the power lines to fall down), and I shudder to imagine the carnage on I-95 and 836 when drivers who can't even avoid accidents during afternoon rainstorms suddenly had to deal with ICE.
A category 1 hurricane making landfall in South Florida is like a "Snow Day" in Cleveland or Buffalo -- work from home today, limp and tipetoe around tomorrow, life as normal on day 3. An afternoon snowstorm that persists into the night would shut down South Florida for almost a week, and probably cause more deaths than a landfalling hurricane.
Errr... you have a point insofar as people more or less ignoring NOAA's own weather reports in preference to commercial alternatives goes, but you're dead wrong about satellites and radar. Those commercial stations depend upon the constellation of satellites and array of radar sites that are operated by NOAA, regardless of whether the actual construction was done by a government employee or contracted out to someone.
Yes, there are a few TV stations that have their own X-band weather radar, but they're mostly eye candy for the local audience (limited range, limited features). For their real forecasting work, they grab the level 2 data from NOAA's radar. You can argue about the need for NOAA's meteorologists, but you'd have to be completely delusional to think their data reconnaissance services could be adequately replaced by commercial alternatives. Not even The Weather Channel has the resources to send out the hurricane hunter planes, let alone maintain its own constellation of satellites and operate an array of radar sites with anything close to the scope and capabilities of NEXRAD and TDWR.
Would a 100% commercial enterprise be more efficient? Certainly. However, efficiency isn't everything. Availability matters, too. It's "efficient" (for the profits of an investor-owned power company) to just accept rolling blackouts for a day or two per year, instead of "over-building" their capacity to make sure it never, ever happens. That doesn't change the fact that it really sucks to be in one of those areas when the blackouts happen, and the cost of coming up with your own backup power is several orders of magnitude more than what it would have cost in higher monthly bills had the network just been engineered to a higher standard in the first place. A private company would roll the dice and risk the loss of a satellite or two for a few years, even though the marginal cost of the spares (spread across ~300 million taxpayers) is next to nothing. A private company can't do that, because it only has a few (compared to 300 million) paying customers, so the extra satellite goes from an extra cent or two per year in taxes to doubled (or more) monthly/annual subscription fees.
Plus, I should probably point out that weather affects interstate commerce in a major way, and is probably one of the most constitutionally-unambiguous responsibilities OF the federal government to handle.
In any case, if you think a spare satellite or two is expensive, take a wild guess how much it would cost to relocate SBX-1 to some location in the middle of the Gulf of Mexico, or 200 miles off the east coast between Jacksonville and Cape Hatteras, to do stand-in duty as the most expensive weather radar platform in the history of meteorology (not to mention the damage it would almost certainly sustain if it were literally put in the path of a half-dozen hurricanes in a single year).
The single hardest thing I've had to do in my house is run wires horizontally. Pulling wires vertically is relatively easy, but fishing them horizontally is a nightmare. Run horizontal conduit across every room, and terminate it at low-voltage frames that face opposite sides of the wall between alternating studs (to avoid having blank wallplates every 2 feet in any one room. If the room backs up against a closet, put frames in the room where you want them, and put the rest of them on the 'closet' side of the wall. That way, if you need to pull wires horizontally in the future, you can just remove the faceplates and pull the cat5/6/fiber/whatever across the room. On each wall, have a junction leading to another conduit that runs to a central wiring area. It's OK and appropriate to have two or more such areas... say, one near the room where your home theater system is located (so you can put most of your gear there, so you can have direct physical access to it from the location where most of your real-world interactive viewing will occur, and feed other rooms semi-directly from there), and at least one on each floor.
I can't emphasize the "have multiple wiring paths and crossover points" enough. I've seem some people go to patently absurd lengths so they could get ideological about having "home runs" for everything... then complain that their system is too brittle, or they had to buy a thousand-dollar HDMI extender instead of a $60 one, because they insisted on sending the signal on a 700-foot detour to some central location instead of just pulling a wire through the wall and feeding the jack on the other side. There's a time and place for both. For networks, go ahead and homerun. For high-bandwidth stuff that has serious, expensive-to-fix problems as the distance increases beyond a hundred feet or so, run point to point. Build your conduit so you can do both. Just because you CAN solve certain problems by throwing expensive hardware at them doesn't mean it's necessarily the best idea if you aren't faced with insurmountable physical constraints. Wires are cheap. Managed VLAN-capable gigabit switches with IGMP 3 snooping aren't.
Ask yourself: if, 7 years from now, you somehow had to get a wire from (some random point in a room) to (some random point in another room on the other side of the house), how would you do it? Could you do it without major surgery as, at worst, a weekend project that might involve cutting a new hole or two for the faceplate frame and some vertical fishing? Because if you can achieve that with conduit, you'll be set for anything you can think of for the next 40 years. Don't even TRY to anticipate the kinds of cables you'll need to run. You'll be wrong, and end up pulling a shit-ton of cable that will end up being sub-optimal for whatever you had in mind, anyway. Run horizontal conduit, interconnect walls with junction points around the house, and sleep well knowing you're ready for anything. I know firsthand -- I spent a fortune, and weeks, pulling shielded cat5e cable after work so I'd be able to run component video over cat5e without interference. I ended up buying $50 HDMI-over-1-cat5e extenders, and all the shielding (god, it was a pain to terminate) ended up being moot.
Remember, high- and low-voltage can't share conduit or boxes (unless divided). I personally recommend running 4 conduits across each wall... 2 for LV, 2 for HV (possibily substituting Romex for one of the HV conduits for the initial build). Use one conduit in each pair for your "phase 1" wiring connecting your "phase 1" boxes, and use the second conduit in each pair to run future wires from "phase 1" boxes to other locations where you can't bring yourself to put blank wallplates right now. By using the second conduit for "the last 2/4/6 feet", you'll minimize the number of wires that have to be pulled out and re-pulled if/when future cuts to condudit #2 are necessary.
For high-voltage wiring where you're putting new boxes, put the deepest boxes you can, with box extenders if possible. If you're putting
Sigh. It's things like that that make me start to seriously think that Google is losing its way, kind of like Microsoft did between NT and Win2k... the point when the original innovators are now too rich to actually be useful as employees anymore, but are politically untouchable & starting to drag everything else down by virtue of their own de-facto disinterest in their official job... delegating everything they can to people whom they refuse to actually give real decision-making authority, then scrambling to cobble things together at the last minute when push comes to shove before dropping off the earth for days/weeks at a time to go sailing, climb Mt. Everest, visit Antarctica, and travel to Baikonur for a few weeks to see whether they have what it takes (besides a few million dollars) to become a Commercial Cosmonaut.
The Nexus One was groundbreaking. The Nexus S was thoroughly 'meh' (admittedly, compounded by the fact that Sprint took so long to get it, by the time it was available, it literally WAS last year's phone), and the Galaxy Nexus was quickly overshadowed by the Galaxy S3, and equally 'meh' by the time Verizon's exclusivity ended.
Arguably, the fact that Google even ALLOWED Verizon to have exclusivity to both the name and franchise, despite having the gall to lock the phone's bootloader and take a corporate dump on everything it means for a device to BE a "Nexus", is more evidence that someone in their Android group has either completely lost touch with reality, or just plain didn't care because he was too busy training for space camp.
I've said it a thousand times, but what Google really needs to do is just pick various manufacturers' best of breed phones for the "Nexus Treatment", and make it so you'd just go and buy a phone like the Galaxy S3, OneX, or RazrMax, then download Google's official AOSP build for it, plug in the USB cable and connect it to your computer, then blow away the carrier's stock ROM with the official Google-blessed AOSP Nexus kernel and distro for it, instead (kind of like buying a laptop with Windows Starter Edition, then blowing it away with Ubuntu or Mint). Otherwise, Google's going to end up having their whole Nexus platform end up turning into "last year's hardware, slightly reheated and limping along with next year's software".
I doubt it, or at least doubt that they'd instantly go into such a mode the moment commercial power were lost. Otherwise, they'd have lots of REALLY angry customers who'd be furious about being unable to call anyone besides 911.
At worst, I could see them load-shedding the fiber or microwave link and falling back to a single T1 if the power situation became dire (or the fiber/microwave link lost its connectivity due to some upstream problem anyway), but prohibiting everything besides 911 would cause them more problems than it would solve. Especially when you consider that during a real hurricane, calling 911 is almost pointless anyway, because they aren't going to risk the lives of police/paramedics/firefighters by sending them out in 100+mph winds with trees and live power lines down on the road.
Still, I guess this means I should also look into buying an old rootable & reflashable GSM phone on eBay while I'm at it. I remember reading somewhere that there was at least one guy working on implementing a v.92 "VOIP-mode" 2400/4800/9600 baud modem for Android that would work by taking direct control over the phone's audio hardware and bitbanging it in realtime, just like a Winmodem would. I think the original intent was to let you do things like send faxes to real-world fax machines straight from your phone, but I don't see why it couldn't be used to connect to a modem bank somewhere, too.
Before someone mentions that modems don't work over cell phones, I should mention that there ARE a couple of modes that were newly defined sometime around v.92 or v94 that are basically high-bandwidth low-bitrate modes that try to trick VoIP or wireless codecs into thinking that they're speech plus loud background noise that can't be separated out or compressed away, and overwhelm the codec into giving up and encoding the content at a higher bitrate than it would ever use for straight 2400 baud.
The idea is that if you try to send straight 2400 baud tones through something like a GSM codec, it'll try to encode them with the equivalent of 1200 bits per second and you'll end up with garbage. But, if you encode 2,400 bits per second along with 48,000 bits per second of Reed-Solomon error correction, and do it in a way that totally overwhelms the compression algorithm, it'll grudgingly end up encoding it with a much higher bitrate, and leave you with enough to recover the original 2400 baud data at the other end. It's mostly used now for fax machines and credit card processing, but in theory it could also be used for straight dialup internet.
What's your exact model number? I'm intrigued, and might want to buy one. I was under the impression that TrippLite's UPSes were among the most intolerant (of frequency deviation) out there.
The most tolerant UPSes I'm aware of are supposedly APC's, IF you explicitly kick them into 'tolerant mode'. I believe they're still rigid about frequency, but will tolerate voltage sags a little longer before kicking in, so it might prevent thrashing with an inverter-type generator.
Side note: only APC's high-end and business-class UPSes officially support 'tolerant mode', but from what I've seen, even their consumer-grade models have the capability... it's just not something you can enable from their official management/monitoring software. However, if you open a serial port and blindly send the right sequence of bytes to the UPS, you can probably enable it on their consumer models, too. I know that I have an ES350, and was able to trigger it just fine. Unfortunately, it didn't fix the problem for me (it made it thrash a tiny bit less, but it still thrashed).