Windows licensing. Officially, new mobo + new CPU == new computer that requires a new OEM Windows license. Old mobo + new CPU == same computer, as long as you reboot it into full stability at least once before (and between) changing the video card & boot drive.
So, it's not *just* $100-200 for a new mobo... it's $100-200 for a new mobo PLUS ~$89-189 for a new Windows license.
For a home user, it's probably moot... Microsoft will let you voice-authorize the serial quite a few times before saying "no"... but for a business large enough to have actual IT staff, license violations are a Big Deal.
Apple won't be happy until they've turned Macbooks into a useless slab of semi-translucent plastic whose only real difference with the future iPad is that you'll be allowed to run apps you've signed with your own key, instead of being restricted to running only apps already-approved and downloaded from Apple's store.
Nothing, except for the fact that many cheap devices still can't *do* 802.11ac & are stuck in a 2.4GHz 802.11n ghetto.
Also, in South Florida (and Washington DC, plus other metro areas) roughly HALF the UNII band is locked out due to nearby weather radar sites using the same frequencies, and most of the rest is stomped-over by goddamn channel-bonding neighbors. Thank GOD channel 165 can't be bonded... it's literally the only one left that still works reliably at my house.
The decision to allow 802.11n 2.4GHz channel bonding (instead of limiting it to 5GHz+) was one of the worst & most inexcusable decisions, ever. I have one neighbor splattering over channels 1-6, another splattering over 3-11, both stomping each other on channels 3-6, and making the entire goddamn 2.4GHz band unusable for me. All I ask for is one fucking low-power narrowband safe haven that works reliably for 10-20mbps & can't penetrate much beyond a single room or two.:-(
Not really. At the end of the day, the radio modem still needs fiber or high-speed copper... the closer to the end user, the better. You will never, ever be able to do the equivalent of stream raw 4k HDMI over 5G in an urban area within a cell larger than a single room, let alone a single-family home or apartment. There just isn't enough spectrum. At gigabit+ speeds, 5G just means you can get away with running a fiber bundle to the curb & distribute it the last thousand feet to outdoor fixed antennas & relay it onwards to indoor wired networks feeding room-sized femtocells. Best-case, your service provider hands you a pile of boxes that "just work" (using whatever crap wiring the house already has) that allows nontechnical users to pretend there's no difference between "the internet" and something their nerdy cousin calls "a LAN".
What I really want to see, though it'll never happen: the FCC partially taking Wifi "channel 14" from Globalstar via eminent domain, then making it legal for Americans to use... but ONLY indoors, with limited power (say, 10mW output, 50mW EIRP) and no channel bonding allowed, so we can have ONE GODDAMN 802.11n channel that neighbors in dense urban areas can't fuck up and ruin(*).
2.4GHz is going to be with us for a long time due to cheap IoT devices, and channel 14 is the only place LEFT in the legacy wifi spectrum that hasn't been ruined by hopeless channel-bonding and neighbors who insist upon (or allow Comcast and AT&T) splattering across channels 1 through 11 with excessive power. ---
(*) Made usable throughout the house by putting something like an old Ubiquiti access point in each room using 5Ghz or ethernet for backhaul. Traditional 802.11n client-initiated handoffs [via 802.11r] don't work well/at all in home environments... devices rarely implement 802.11r properly, and most consumer wifi gear is completely oblivious to it. Ubiquiti moved the logic to the access points... they all compare signal-strength notes, and mutually spoof each other to trick dumb clients into connecting to the best one anyway. I think Ubiquiti took away that feature last year for some crazy reason, but I like to hope they'd bring it back if they had a compelling reason to do so.
Modern-day Mexico City occupies quite a bit more land area than the site of that small, formerly-marshy island, the same way modern-day "Los Angeles" is unfathomably larger than the Spanish-Mexican settlement of "El Pueblo de Nuestra Señora la Reina de los Ángeles de Porciúncula". As others have noted, the problem isn't that Mexico or Mexico City lacks water, the problem is inadequately-sized (or maintained) infrastructure.
In Mexico City's defense... it's not always a matter of poverty or corruption, either. Water and sewer mains tend to get built beneath roads... roads that often grow into 6-8 lane divided highways by the time the original mains need to be replaced and upgraded. One does not simply close a major arterial highway for a year and a half without MAJOR political repercussions, because the life of everyone who lives nearby is going to SUCK MISERABLY for months... maybe years.
Complicating the equation even more, the direct beneficiaries of that reconstruction usually live quite some distance away from the area where the most disruptive construction is required (as the city sprawls into new areas, new water & sewer mains get built with abundant capacity to absorb future expected growth... but ultimately feed from (or dump into) the same older mains through the city's original neighborhoods that are the ones that will need large-scale reconstruction).
Water & sewer mains aren't cool or sexy, but they're essential municipal infrastructure that costs an ENORMOUS amount of money, and takes DECADES to plan, fund, and build. In most cases, by the time a city or county realizes it has a problem, it's ALREADY way too late to do anything about it before the problem becomes even worse.
As someone mentioned earlier, the neighborhood in question is one that experienced unanticipated explosive growth that completely outstripped earlier city planning efforts. In this case, it was due to earthquakes... after a major disaster in a big city ANYWHERE, it's common to see a HUGE surge of new construction in nearby "greenfield" areas, because it's a lot faster (and cheaper) to build vast tracts of new homes on virgin land than it is to try "redeveloping in place". Redevelopment of existing neighborhoods is expensive and slow to begin with, and doing it in destroyed neighborhoods after a disaster is even worse. Let's suppose Developer #1 decides to rebuild a destroyed neighborhood, while developers #2-10 quickly build sprawling new suburbia across formerly-rural greenfield. By the time developer #1 finishes his first new homes, the market will be hopelessly saturated with a glut of new, cheaper homes built by the other 9 developers. Most of the time, the first new homes built on the greenfield sites will be ready to sell before the developer rebuilding the old area even manages to get the destroyed area cleared and READY to build on.
I'll believe Mattel is serious when half the 'Barbie' aisle looks like the 'Maker' department at a store like Fry's. Say, Barbie-themed RasPi & Arduino boards, cases, and tools, plus the usual components & accessories.
> Why should there be ANY cost attached when the computer already has a legitimate license that has already been paid in full?
In most cases like this, Windows has an enterprise-licensed copy of windows that's bound to both the original computer AND licensee & can't be transferred to another user, EVEN IF it's the original installation with the original hard drive.
If Windows is installed from a literal retail-licensed copy of Windows, you can officially reuse the license on new computers in perpetuity, as long as you practice strict "serial monogamy" & never have multiple computers capable of using that license at any one time (and especially never have 2+ computers using the same license online simultaneously -- the BIG "no-no" that Microsoft notices). In practice, VERY FEW actual retail-licensed copies of Windows exist because they're SO EXPENSIVE compared to OEM-licensed copies.
In theory, a business could buy & resell retail windows licenses, but it would be legal suicide since you (the reseller) could never truly guarantee the seller actually complied with the terms & destroyed all other copies & installations of it.
OEM-licensed Windows is black & white on paper, but gray & blurry in practice. Officially, if a computer is sold by an official volume reseller like Dell with Windows, it's unambiguously legal to reinstall windows from the original hard drive recovery image using the original serial... but in practice, the original hard drive is usually long gone (or in a state whose integrity from malware can't be guaranteed). All other commercial resale scenarios probably violate Microsoft's licensing terms.
If you're an individual, Microsoft doesn't particularly care. They'll unofficially tolerate quite a few things that are officially forbidden. Reinstall an OEM copy of Windows on 3 or 4 different computers, and they're unlikely to do much beyond make you phone in to activate it (subtly letting you know, "We're watching you") as long as their servers never see a computer with an older-activated copy of that license online after you've activated it on a newer computer. ESPECIALLY if you're just using an old OEM copy to do a clean installation of Windows on a new computer that shipped with its own OEM license for the same version of Windows ANYWAY.
TL/DR: just because a once-valid Windows license code WORKS to (re)install Windows doesn't necessarily mean it's officially legit NOW in Microsoft's eyes... especially if you're doing it commercially and reselling computers.
Not really. For typical desktop use running apps that don't necessarily TRY to be optimized for multithreading, Windows generally "tries harder" to put multiple cores to good use ANYWAY.
The traditional Linux attitude has been, "if you want to take advantage of multiple cores, write your software properly." Windows just assumes (correctly) that most apps DON'T try, and does its best to at least multithread the *rest* of the system (including libraries). It doesn't always succeed, but it tends to get accidentally-good results more often than Linux does.
Where Windows USED to fuck up badly was memory usage. Simply put, older versions were WAY too eager to swap out to virtual RAM, even if you had 16+ gigabytes. As a practical matter, without registry-tweaking in semi-undocumented ways, it was almost IMPOSSIBLE to get Windows to use more than a gig of RAM for any one app, or 4 gigs system-wide, without having Windows start aggressively paging out virtual memory, unless you disabled paging altogether.
I believe it happened because the logic for virtual-memory handling was intertwined with the logic for implementing PAE. Prior to ~Win8, there were two operating modes: one for low-memory systems whose history went all the way back to Windows95 (and allowed it to limp with a mere 4mb of ram), and one for high-memory systems where PAE was assumed to be used (even though Microsoft ultimately disabled PAE in most systems).
When 64-bit Windows arrived, Microsoft started with the "high-memory" code, ripped out the PAE-implementation logic, but ran out of time/developers and left the PAE-imposed LIMITS in place. It wasn't until ~2014 that Microsoft finally got around to refactoring it to remove those limits. And even now, many of those limits still exist by default unless you disable them via registry keys because some popular older apps crash if they're disabled.
Don't believe me? Do a virgin-install of 64-bit Windows 10 on a new PC with 32gb of RAM, then TRY to get Windows to physically use more than 4 gigs of it without swapping (and without registry-hacking or disabling virtual memory altogether). What Windows NOW does by default is to swap pages to UPPER ram, with slightly better logic to decide when it needs to swap those "paged pages" to the ssd/hard drive. I think it also now can do most of it via register-manipulation instead of copying. But at the end of the day, Windows STILL defaults to treating your glorious 32gb ocean of ram as a glorified ramdisk unless you force it to do otherwise.
In case anybody's wondering why the mode is disabled for laptops, it's because quite a few literally *can't* run for extended periods of time at full-bore speed without overheating. For YEARS, companies like Toshiba were advertising specs based on CPU specs, but underclocking them behind the scenes to reduce heat. Most current laptops can (mostly) run at full-speed without crashing, but will gradually cook their electrolytic capacitors & other components to death if you insist.
Now, the big question is... what's Microsoft's definition of "laptop", and how does Windows decide? Is a mini-ITX "desktop" with mobile i5 a desktop or a laptop? How about a Clevo-made gaming beast with a "desktop" i7 & discrete (but semi-proprietary) graphics card that -- electronically -- is still PCIe (with nonstandard form factor)?
I wouldn't mind paying someone like Amazon or Google $10/month for access to every meaningful newspaper in America (with Google dividing it up among the papers I read that month), but I refuse to get sucked into a half-dozen monthly subscriptions... especially when seemingly all of them are "pay {some reasonable} rate for the first {n} weeks, then {get ass-raped} thereafter until you notice and cancel". I MIGHT do it if there were an option to automatically end the subscription once the promo rate expires, but over the past few years, I've gotten to the point where I automatically tell anyone trying to get me to sign up for teaser rates that silently go up to just go fuck themselves and die. I fell for subscription scams like that all the time when I was younger, but now it just seems like total bullshit and I refuse to put up with it anymore.
You can blame Verizon for plenty of shitty things, but CDMA2000 isn't one of them.
Verizon Wireless began its life outside the NY metro region (and probably WITHIN it) as "PrimeCo" -- a company that was "CDMA" from day one. At the time PrimeCo launched in the mid-90s, it had four real-world alternatives:
* Analog. Never a real alternative, but listed for comprehensiveness.
* 1G GSM. Nearly-irrelevant outside of Europe. Almost by definition, mid-90s 1G GSM mandated frequencies that would have been impossible to license in North America. Even IF a US carrier used "GSM" with "North American" frequencies, the likelihood that a non-North-American "GSM" phone would have been interoperable was approximately "none". Compatible dual-band phones eventually appeared, but were the overwhelming exception rather than the rule until well into the early 2000s.
* TDMA. In the US, used primarily by AT&T. Basically, AT&T skimmed off the low-level parts of 1G-GSM, lopped off the parts that were needed to permit interoperability with other networks, and tweaked the standard in ways that allowed them to exceed the 35km max tower-phone distance limit that was baked into GSM.
* IDEN. Was basically another semi-proprietary flavor of TDMA whose main selling point was its ability to handle point-to-point communications among employees all connected to the same tower without incurring airtime charges. Popular with businesses, but even in its heyday, nobody expected it to have a long-term future in the US.
Compared to CDMA, both 1G-GSM and TDMA were archaic & primitive. CDMA is SO advanced (compared to TDMA, which GSM and IDEN were both flavors of), it almost looks like pure black magic:
* Soft hand-offs -- switch from tower to tower without anybody noticing. With GSM & TDMA, you heard a distinct click as you were handed off.
* Few hard constraints about tower placement and spectrum allocation. With GSM & TDMA, tower location & spectrum allocation has to be carefully planned, and reconfiguring later is a Very Big Deal. With CDMA, you can literally solve congestion problems by semi-randomly throwing new towers into a congested area & the network will literally "fix itself" (that's not to say good planning is irrelevant... good planning gets better and more cost-effective results... but with GSM/TDMA, it's absolutely 100% mandatory and non-negotiably REQUIRED).
* More efficient spectrum reuse. Simply put, you can transmit more total bits among all active users over the span of a second using a finite chunk of spectrum using CDMA than TDMA/GSM/IDEN.
The bigger clusterfuck in the US happened due to spectrum chaos and Qualcomm's certification policies. In theory, almost any top-shelf Android device sold since ~2013 is technically capable of doing 3G GSM and CDMA2000-EVDO on any network in America... except they aren't. Sprint's network won't allow customers to "natively" use any phone not literally sold by Sprint (eg, Sprint will allow one of the few remaining Canadian Telus CDMA2000 phones to roam on Sprint, but if the owner moves to the US, Sprint won't allow them to sign up for service using the phone that roamed perfectly well on the same network). Likewise, Verizon will grudgingly allow you to use an unlocked CDMA2000 phone on their network... but they won't lift a finger to give you the SLIGHTEST bit of help, and a non-Verizon CDMA2000 phone will NEVER be able to use EVDO on Verizon due to the way they implemented authentication (only CDMA2000-1xRTT, which is about 150kbps). An unlocked AT&T phone might be WILLING to do 3.5G HSPA on T-Mobile with a T-mobile SIM card, but only on frequencies used (or formerly used) by AT&T. An unlocked T-Mobile[US] phone might be willing and able to do 3.5G HSPA on AT&T with an AT&T SIM card, but nevertheless be unable to do LTE at all REGARDLESS of frequency due to Qualcomm's licensing policies.
Qualcomm's licensing policies deserve special note. Basically, Qualcomm licenses radio modem firmware to CARRI
Musk's brilliance was having the Falcon's rockets designed for vertical soft landings. Among other things, this makes a certain degree of failure-risk during launch more acceptable.
With conventional rockets, rocket failure (of almost any kind) during launch almost inevitably meant payload loss. If you were *lucky*, you might be able to make a last-ditch effort to get your satellite into some kind of random orbit that might have value to a buyer SOMEWHERE, as opposed to writing it off as a total loss from day one.
With Falcon, if something goes wrong during launch & SpaceX realizes that they won't be able to get a satellite payload into its intended orbit, they could conceivably abort the launch, bring the whole thing back to Earth for a soft landing, and make another attempt to launch the same satellite another day. That's HUGE.
If nothing else, it means there's now an incentive to bring expensive satellites from failed launches back to earth, instead of making a last-ditch attempt to launch them into some random orbit anyway on the slim chance they'll end up being worth something to someone & not just end up as more space junk. It also means many customers will be able to roll the dice and skip buying a second satellite as a launch-spare, since "launch failure" won't necessarily mean "satellite loss".
In effect, SpaceX has potentially reduced costs at BOTH ends... rockets and launches themselves are cheaper, and the consequences of a "softly" failed launch are potentially reduced as well.
Bitcoin's fundamental problem is that if it doesn't crash & lose its value anyway, it'll eventually become too valuable to use as actual currency.
Bitcoin's proponents point to its decentralized nature and lack of government control, but overlook the fact that its blockchain can't scale. It's ALREADY staggeringly huge (something like 200 gigabytes the last time I checked). Sure, you can delegate responsibility for checking it to some other device (or third party) instead of hauling the whole thing around on your phone, but THEN you become dependent upon your ability to trust that third party AND its ability to keep bad guys from hacking them (to show that a spent bitcoin is still available, long after it has ceased to be). The second part is what fucks you if you try using YOUR OWN off-device infrastructure (like a server at home) or run by some public-spirited organization (that nevertheless assumes zero liability for anything that Goes Wrong)... not even large banks that actively TRY to prevent hacking attacks are able to succeed with only technical defenses. The entire credit card industry works because they have enough capital to absorb losses & use the legal systems of their respective countries to keep crime down to a tolerable slow burn. Your personal long-term ability to keep your own server hackproof (while nevertheless keeping it accessible to you when you're away from home) is approximately zero... eventually, someone will find a way to compromise your phone, your server, the PKI it depends upon, and/or anything in between... unless (or quite probably, EVEN IF) you dedicate your life to maintaining its security and integrity.
And in the meantime, you're going to spend more and more time waiting for transactions to clear unless you pay someone to expedite the transaction and assume at least some risk.
So, we get to problem #2... Bitcoin transactions can't be added to the blockchain until someone else manages to mine a new Bitcoin. As time goes on, this will become exponentially more expensive and difficult, so the fees they'll charge will go up & overwhelmingly impact otherwise-small transactions. If the cost of completing a transaction approaches something like 0.1 Bitcoin, it'll cease to be cost-effective to trade anything less than several Bitcoins at once. At best, Bitcoin becomes a virtual backing currency that only large traders can afford to exchange. To an extent, this has already happened with companies that use commercial wallets to handle microtransactions. And the companies running those microtransaction wallets are just as vulnerable as everyone else, except they're even BIGGER targets and even MORE hopelessly outgunned (think: homeowner vs armed burglar, compared to small business vs organized crime shakedown).
Long-term, Bitcoin will probably survive and pivot to new primary users when others fall, but it's hopelessly naive to think Bitcoin will EVER be usable for casual purchases. Commercial services that absorb risk to expedite transactions will end up costing as much as Visa or Mastercard, and be every bit as regulated by governments. As an almost-free guerrilla alternative to Paypal, Bitcoin is a long-term failure by design. As a virtual reserve currency, it might survive... but whether Bitcoin will ever become more cost-effective for businesses and banks to use for real trade than US Dollars, Euros, or any other currency is far from certain.
> (I'm not sure why one would want pages in a text heavy book).
Historically, technical books follow the pattern of, "Display a diagram, chart, or photo on one page, and talk about it on the facing page". Reading such books is MUCH harder when you can only see one page at a time, and trying to do it on a device that takes hundreds of milliseconds to flip between pages will slowly drive you mad. Slow page-flipping times is the #1 reason why reading technical books in ebook form is such an awful experience.
For an example of how fast eBook readers NEED to be able to render in order to not completely suck, check out this video: https://www.youtube.com/watch?...
I'm pretty sure the guy who made the video pre-rendered the entire book to bitmap images, because I doubt whether ANY tablet available in 2012 could have rendered pages from a typical PDF textbook quickly enough to render every page from scratch in realtime.
Was NASA *seriously* planning to launch a shuttle per week using the fleet of shuttles we had, or were they planning to build a much, MUCH larger fleet?
Considering how much large-scale refurbishment/remanufacturing the shuttle itself ended up needing after every launch (esp. the heat shield tiles in later years), I can't see how they could have even FANTASIZED about sustaining that kind of launch schedule with such a small fleet.
Did they strip out the Tesla's main battery (to reduce weight), or is that what's powering the transmitter & camera? It seems like a fully-charged Tesla battery capable of driving 200+ miles at 80mph SHOULD be more than capable of powering a camera and 100-watt transmitter for quite a while (100 watts doesn't sound like a lot, but it's actually the same amount of power used by the Mars Reconnaissance Orbiter when it uploads image data from the rovers back to NASA via the Deep Space Network).
For ebook-reading, 4:3 is actually a superior aspect ratio for most content. If you're scaling to fit 1 page on the screen, a 9" 4:3 is like a 10-11" 16:9.
For ebook-reading TWO pages side by side, a landscape 3:2 aspect ratio is ideal, because it's like two 4:3 (3:4) portrait displays side by side.
In any case, high-dpi (200ppi or better) is important... reading text on a 18" 1920x1080 display is like trying to read 1024x768 text on a shit 15" Packard-Bell CRT with.39mm dot pitch circa 1998. For all intents and purposes, a 240ppi OLED is more crisp for text than a 300dpi laser printer with premium paper (because it's 240ppi 8-16 bit grayscale, vs 300dpi 1-bit black/white).
The main problem with tablet-as-ebook-reader: they need bigger high-dpi screens (13-14" would be nice) AND fast primary storage & cpu. A tablet as aa ebook reader might spend 99% of its time doing nothing... but when you go to flip pages, it needs 10-20ms render time, MAX. Longer might be ok for things read serially (like fiction), but current ebook solutions SUCK for random-access reading, like technical manuals.
The Chuwi Hi12 has a good 12.1" 2160x1440 display at a fairly sane price... but it's an older model, and its other specs aren't much to write home about (mediocre wifi, average battery life, and slow microUSB charging via a port that doesn't inspire much confidence in its likely longevity.
That said, it's decent for 2-up ebook reading (a bit on the heavy side, but not unbearable).
No, YOU'RE the one who's full of shit. Lego Legends of Chima isn't playable on a factory-fresh 3DS until you allow the 3DS to fetch some mandatory update from Nintendo. It doesn't require continuous online connectivity to play it, but until the update is done, you won't be allowed to play the game. If you had a new-old-stock shrinkwrapped 3DS that spent 30 years in a closet and tried to play that game for the first time sometime around the mid 2050s, you wouldn't be allowed to do it.
As far as I can tell, Nintendo has an incrementing counter for patch level, and 3DS cartridges can specify a minimum patch level. If you put the cartridge in a 3DS that hasn't been patched to that level, the game won't run until the 3DS finds some way to connect to the internet, download the patch, and install it. Thus, if you're attempting to play a game for the first time that was released late in the 3DS's official life and requires a patch level that's higher than what that new-old-stock 3DS shipped with from the factory (or even a non-shrinkwrapped old 3DS that never managed to get updated to a newer patch level), you aren't going to be able to EVER play that game without some major hacking. Given Nintendo's past behavior, the likelihood that they'll have a server up and running to handle 3DS patch downloads circa 2050 is somewhere between "slim" and "none"... as is the likelihood that Nintendo (or whomever owns their copyrights at that point) would tolerate allowing anyone to independently distribute their old firmware updates after they themselves have discontinued support for them.
So fuck you. If you think this is "fake news", you're probably dumb enough to BE a prime target audience member FOR "fake news".
In the past, consoles either had no system software or shared libraries at all, or publishers were expected to distribute their games along with any patches to the baseline system software and libraries that their games depended upon. It wasn't until consoles gained network connectivity "out of the box" that companies felt entitled to start making them dependent upon post-manufacturing updates that could be made unavailable at any time without recourse from consumers.
IMHO, the manufacturer of any device that's unfit for its advertised purpose without additional downloads should be required by law to provide SOME "reasonable" way for a knowledgeable (not necessarily technically-clueless) end user to obtain and apply those updates on his own, LONG after the vendor itself has ceased to provide them online. For example, suppose a scenario like...
1. Hypothetical IETF RFC for "Post-EOL Firmware Update Service" that describes a protocol for a service you can run on a TCP/IP network and serve a firmware update to a bootloader long after a device has ceased to be supported by the original manufacturer.
2. A vendor like Nintendo escrows its firmware updates with the Library of Congress, and pays some nominal fee like $100 to register them. Ditto, for any software vendor who ships software that can't actually run without some mandatory update that isn't shipped along with it. Every 30-90 days, the LoC publishes the escrowed code and sells it to distributors who can resell it, give it away, or otherwise do whatever they want to with it. The LoC also maintains an index of escrowed code.
3. Years later, someone buys an antique 3DS at a flea market, along with some old games, then discovers he needs a patch to allow them to work. After doing some research, he learns that he needs to update that 3DS to patch #24.8.17.9 or newer. He could technically order it from the Library of Congress for $850, but he can instead get it free from "ancientfirmwareupdates.com" in return for watching 15 minutes of ads (because companies would buy and archive that published data, then redistribute it themselves with the LoC's full blessing). He sets up a server that implements the required service, points it to the patch data, and sets up an equally-antique 802.11n access point with SSID "3DSUPDATE". After consulting
And for approximately $25, you can buy a USB3.0 BD-ROM + DVD+/-RW drive the approximate size of a CD jewelbox that can also do CD-R(W).
You don't buy CDs to listen to directly... you buy them to get a clean source that can't be arbitrarily taken away from you {n} years from now because some company decided that "for life" means "the life of our product, as we define it" (e.g., Zune), shuts down their DRM servers, and leaves you with either nothing at all, or (at best) one final, fragile copy that'll be gone forever when the shit electrolytic capacitors (or glued-in battery) dies 2 or 3 years later.
This past Christmas, I spent a day playing with my old C64 & Vic-20. Both worked perfectly, and so did my old 1702 monitor. My old Odyssey2 (Videopac, in Europe) and RCA Studio II worked, too. It was a sobering experience when it sank in that there's probably not a single goddamn computer or game console you could buy today, put (shrinkwrapped) in a closet for ~35 years, and have ANY reasonable hope that it will actually WORK (and be usable "for real") when you power it up for the first time. Even a goddamn Nintendo 3DS refuses to let you do anything with most games until it connects to the internet & updates... once Nintendo shuts the servers down someday (like they're doing with the Wii's e-store this year), bye-bye system. Try doing anything useful with a Logitech Revue, a Zune, or a "WMV-HD" disc from ~10 years ago (Microsoft shut down their DRM servers & said 'fuck you' to customers, so they're now unplayable). This is why I'll NEVER spend money on full-priced content tied to Microsoft devices or services again, and why Google will probably never convince me to take their future media devices seriously. I'll throw down $7.99 for a used game, maybe, but never more than $29.99 (very rarely, more than $19.99), because I now assume anything I buy will be taken away 3-5 years from now & devalue it accordingly.
I can think of one area where reduced costs of long-haul trucking could result in exponentially greater utilization by consumers: storage. Especially if electric trucks had enough battery capacity to recharge in places where electricity is cheaper than nearby areas. Right now, we already HAVE "cube" storage... but the cubes themselves are still stored somewhere semi-nearby. If the cost of trucking those cubes 250 miles were only slightly greater than the cost of trucking those cubes 20 miles, instead of storing those cubes in warehouses that are only slightly cheaper than the area where they came from (say, New Jersey vs Manhattan), you can send them to a vast, sprawling, robotic storage facility out in the middle of rural Pennsylvania. It's kind of a contrived example, but shows how if something becomes cheap and easy enough, brute-force solutions that would be cost-prohibitive NOW can become financially viable.
The CODEC is another potential snag. Lossy algorithms like MP3 make a POINT of mangling and filtering away things that are "inaudible", especially if there are lots of OTHER things going on in the audio at that instant.
I'm not sure about AC3 & Dolby Digital, but I know that MP3 absolutely DESTROYS phase relationships used by algorithms like Qsound (for a very audible example, find a CD of Madonna: The Immaculate Collection, rip it, compress it to MP3 at 384kbps, 256kbps, 128kbps, and 64kbps (all CBR, not VBR), then compare them all to the original.
In the uncompressed CD audio, it feels like you can reach out and touch the audio.
At 384kbps, the spatiality is gone, but the stereo separation and frequency response are basically intact.
At 256kbps, the stereo separation "comes and goes"... mostly intact, but occasionally flattening out.
At 128kbps, the stereo separation is practically GONE, and you can hear the high-frequency response falling away.
At 64kbps, you might as well be listening to a cheap 1970s transistor radio (FM) with a 3" speaker.
I'd expect that the effect might actually be WORSE with bitrate-starved Dolby Digital Surround and DDS, because having multiple discrete surround channels partially hides the effect of mangling stereo phase relationships. The compression algorithm might tolerate and preserve higher-frequency audio on certain channels, but not others (e.g., if you mixed a 19KHz sine wave into the center channel or surround channels, it might be attenuated away, but might be at least partially preserved if you mixed it into the front left or right channels).
This is part of the reason why Dolby ProLogic Surround (note1) works somewhat well with audio from high-bitrate sources, but falls flat on its face & dies with many compressed-audio sources. For example, try watching a Hollywood Movie streamed via SlingTV with your AVR set to decode it as ProLogic surround sound. On many channels, the audio will just completely fall apart, generating endless pops and other audio artifacts on the surround channels unless you forcibly disable ProLogic and make the AVR treat it like plain stereo (most AVRs will default to treating 2-channel PCM [streamed as AC3, decoded to PCM2.0, and sent over HDMI by a Roku] as de-facto ProLogic, because often the 2-channel stereo track IS surround-encoded-as-ProLogic. SlingTV's audio bitrate is too low, and mangles it into oblivion.)
TL/DR: high-frequency audio is one of the first things that gets kicked to the curb (or mangled) by bitrate-starved surround-sound compression algorithms.
-----
note 1: Back in the late 1970s, Dolby came up with "Dolby Surround" for theaters. Broadly speaking, it allows you to encode a low-fidelity third (rear surround) channel into the left and right stereo signals in a way that makes it sound like reverb if played back by stereo gear. The catch is, that rear surround channel can't deviate from the audio from the left and right channels by more than a few dB, or it will audibly "leak" into the front speakers.
ProLogic takes the "Dolby Surround" signal, decodes the third channel, then runs everything through a DSP to further transform it into something that (vaguely) resembles 5.1 surround. It's a pale imitation, but when done well, can sound better than plain stereo.
Because analog Dolby Surround (and ProLogic surround) is mastered into the analog stereo soundtrack of a movie, a lot of movies distributed on LaserDisc and VHS ended up having passable surround sound by default if you ran the stereo audio through a surround sound processor. It was basically a lucky accident, but it worked.
DVDs arrived, and digital surround (primarily Dolby Digital 5.1 and DDS 5.1) came along with it, and ProLogic became something that was "there" by default in the stereo soundtrack (mostly, because once you mixed it for 5.1, downmixing it to ProLogic was something any decent software could do automatically). Blu Ray took this and extended it further to 7.1-channel DD+ and
Exactly. Long before ANY colonists arrive, the terrain surrounding that first colony will be LITTERED with hundreds of earlier landing craft that delivered robots to begin building & running its infrastructure. There will be THOUSANDS of virtual colonists who've spent YEARS vicariously "living" on Mars by overseeing those robots remotely from Earth (obviously, not in realtime, but spending hours per day reviewing "their" robot's past Sol). By that point, there will even be three sources of recyclable resources:
1) a distant area where the earth-mars rockets used to transport skycrane-deliverable robots were crashed.
2) a nearby area where the descent vehicles were crashed after delivering robots in a skycrane-like manner, and where cargo craft landed normally.
3) All the robots that quit working after arrival.
Sending ANY humans before the robots have built suitable living & working quarters & have things like life support & power generation working flawlessly would be stupid & irresponsible.
You're assuming that whomever started a global thermonuclear war (or their target) wouldn't have a missile or two aimed at Mars, just to take out the enemy's colonies, too.
Imagine it's 1980 in a parallel universe where only the Soviet Union had a colony on Mars. Do you really think the US military wouldn't have classified it as a major target & had at least 1 robotic rocket in its arsenal capable of flying to Mars & launching one or more missiles at their colonies from space? It's sad, but it's totally naive & unrealistic to think otherwise.
Slashdot Mirror
Windows licensing. Officially, new mobo + new CPU == new computer that requires a new OEM Windows license. Old mobo + new CPU == same computer, as long as you reboot it into full stability at least once before (and between) changing the video card & boot drive.
So, it's not *just* $100-200 for a new mobo... it's $100-200 for a new mobo PLUS ~$89-189 for a new Windows license.
For a home user, it's probably moot... Microsoft will let you voice-authorize the serial quite a few times before saying "no"... but for a business large enough to have actual IT staff, license violations are a Big Deal.
Apple won't be happy until they've turned Macbooks into a useless slab of semi-translucent plastic whose only real difference with the future iPad is that you'll be allowed to run apps you've signed with your own key, instead of being restricted to running only apps already-approved and downloaded from Apple's store.
Nothing, except for the fact that many cheap devices still can't *do* 802.11ac & are stuck in a 2.4GHz 802.11n ghetto.
Also, in South Florida (and Washington DC, plus other metro areas) roughly HALF the UNII band is locked out due to nearby weather radar sites using the same frequencies, and most of the rest is stomped-over by goddamn channel-bonding neighbors. Thank GOD channel 165 can't be bonded... it's literally the only one left that still works reliably at my house.
The decision to allow 802.11n 2.4GHz channel bonding (instead of limiting it to 5GHz+) was one of the worst & most inexcusable decisions, ever. I have one neighbor splattering over channels 1-6, another splattering over 3-11, both stomping each other on channels 3-6, and making the entire goddamn 2.4GHz band unusable for me. All I ask for is one fucking low-power narrowband safe haven that works reliably for 10-20mbps & can't penetrate much beyond a single room or two. :-(
Not really. At the end of the day, the radio modem still needs fiber or high-speed copper... the closer to the end user, the better. You will never, ever be able to do the equivalent of stream raw 4k HDMI over 5G in an urban area within a cell larger than a single room, let alone a single-family home or apartment. There just isn't enough spectrum. At gigabit+ speeds, 5G just means you can get away with running a fiber bundle to the curb & distribute it the last thousand feet to outdoor fixed antennas & relay it onwards to indoor wired networks feeding room-sized femtocells. Best-case, your service provider hands you a pile of boxes that "just work" (using whatever crap wiring the house already has) that allows nontechnical users to pretend there's no difference between "the internet" and something their nerdy cousin calls "a LAN".
What I really want to see, though it'll never happen: the FCC partially taking Wifi "channel 14" from Globalstar via eminent domain, then making it legal for Americans to use... but ONLY indoors, with limited power (say, 10mW output, 50mW EIRP) and no channel bonding allowed, so we can have ONE GODDAMN 802.11n channel that neighbors in dense urban areas can't fuck up and ruin(*).
2.4GHz is going to be with us for a long time due to cheap IoT devices, and channel 14 is the only place LEFT in the legacy wifi spectrum that hasn't been ruined by hopeless channel-bonding and neighbors who insist upon (or allow Comcast and AT&T) splattering across channels 1 through 11 with excessive power.
---
(*) Made usable throughout the house by putting something like an old Ubiquiti access point in each room using 5Ghz or ethernet for backhaul. Traditional 802.11n client-initiated handoffs [via 802.11r] don't work well/at all in home environments... devices rarely implement 802.11r properly, and most consumer wifi gear is completely oblivious to it. Ubiquiti moved the logic to the access points... they all compare signal-strength notes, and mutually spoof each other to trick dumb clients into connecting to the best one anyway. I think Ubiquiti took away that feature last year for some crazy reason, but I like to hope they'd bring it back if they had a compelling reason to do so.
Modern-day Mexico City occupies quite a bit more land area than the site of that small, formerly-marshy island, the same way modern-day "Los Angeles" is unfathomably larger than the Spanish-Mexican settlement of "El Pueblo de Nuestra Señora la Reina de los Ángeles de Porciúncula". As others have noted, the problem isn't that Mexico or Mexico City lacks water, the problem is inadequately-sized (or maintained) infrastructure.
In Mexico City's defense... it's not always a matter of poverty or corruption, either. Water and sewer mains tend to get built beneath roads... roads that often grow into 6-8 lane divided highways by the time the original mains need to be replaced and upgraded. One does not simply close a major arterial highway for a year and a half without MAJOR political repercussions, because the life of everyone who lives nearby is going to SUCK MISERABLY for months... maybe years.
Complicating the equation even more, the direct beneficiaries of that reconstruction usually live quite some distance away from the area where the most disruptive construction is required (as the city sprawls into new areas, new water & sewer mains get built with abundant capacity to absorb future expected growth... but ultimately feed from (or dump into) the same older mains through the city's original neighborhoods that are the ones that will need large-scale reconstruction).
Water & sewer mains aren't cool or sexy, but they're essential municipal infrastructure that costs an ENORMOUS amount of money, and takes DECADES to plan, fund, and build. In most cases, by the time a city or county realizes it has a problem, it's ALREADY way too late to do anything about it before the problem becomes even worse.
As someone mentioned earlier, the neighborhood in question is one that experienced unanticipated explosive growth that completely outstripped earlier city planning efforts. In this case, it was due to earthquakes... after a major disaster in a big city ANYWHERE, it's common to see a HUGE surge of new construction in nearby "greenfield" areas, because it's a lot faster (and cheaper) to build vast tracts of new homes on virgin land than it is to try "redeveloping in place". Redevelopment of existing neighborhoods is expensive and slow to begin with, and doing it in destroyed neighborhoods after a disaster is even worse. Let's suppose Developer #1 decides to rebuild a destroyed neighborhood, while developers #2-10 quickly build sprawling new suburbia across formerly-rural greenfield. By the time developer #1 finishes his first new homes, the market will be hopelessly saturated with a glut of new, cheaper homes built by the other 9 developers. Most of the time, the first new homes built on the greenfield sites will be ready to sell before the developer rebuilding the old area even manages to get the destroyed area cleared and READY to build on.
I'll believe Mattel is serious when half the 'Barbie' aisle looks like the 'Maker' department at a store like Fry's. Say, Barbie-themed RasPi & Arduino boards, cases, and tools, plus the usual components & accessories.
> Why should there be ANY cost attached when the computer already has a legitimate license that has already been paid in full?
In most cases like this, Windows has an enterprise-licensed copy of windows that's bound to both the original computer AND licensee & can't be transferred to another user, EVEN IF it's the original installation with the original hard drive.
If Windows is installed from a literal retail-licensed copy of Windows, you can officially reuse the license on new computers in perpetuity, as long as you practice strict "serial monogamy" & never have multiple computers capable of using that license at any one time (and especially never have 2+ computers using the same license online simultaneously -- the BIG "no-no" that Microsoft notices). In practice, VERY FEW actual retail-licensed copies of Windows exist because they're SO EXPENSIVE compared to OEM-licensed copies.
In theory, a business could buy & resell retail windows licenses, but it would be legal suicide since you (the reseller) could never truly guarantee the seller actually complied with the terms & destroyed all other copies & installations of it.
OEM-licensed Windows is black & white on paper, but gray & blurry in practice. Officially, if a computer is sold by an official volume reseller like Dell with Windows, it's unambiguously legal to reinstall windows from the original hard drive recovery image using the original serial... but in practice, the original hard drive is usually long gone (or in a state whose integrity from malware can't be guaranteed). All other commercial resale scenarios probably violate Microsoft's licensing terms.
If you're an individual, Microsoft doesn't particularly care. They'll unofficially tolerate quite a few things that are officially forbidden. Reinstall an OEM copy of Windows on 3 or 4 different computers, and they're unlikely to do much beyond make you phone in to activate it (subtly letting you know, "We're watching you") as long as their servers never see a computer with an older-activated copy of that license online after you've activated it on a newer computer. ESPECIALLY if you're just using an old OEM copy to do a clean installation of Windows on a new computer that shipped with its own OEM license for the same version of Windows ANYWAY.
TL/DR: just because a once-valid Windows license code WORKS to (re)install Windows doesn't necessarily mean it's officially legit NOW in Microsoft's eyes... especially if you're doing it commercially and reselling computers.
Not really. For typical desktop use running apps that don't necessarily TRY to be optimized for multithreading, Windows generally "tries harder" to put multiple cores to good use ANYWAY.
The traditional Linux attitude has been, "if you want to take advantage of multiple cores, write your software properly." Windows just assumes (correctly) that most apps DON'T try, and does its best to at least multithread the *rest* of the system (including libraries). It doesn't always succeed, but it tends to get accidentally-good results more often than Linux does.
Where Windows USED to fuck up badly was memory usage. Simply put, older versions were WAY too eager to swap out to virtual RAM, even if you had 16+ gigabytes. As a practical matter, without registry-tweaking in semi-undocumented ways, it was almost IMPOSSIBLE to get Windows to use more than a gig of RAM for any one app, or 4 gigs system-wide, without having Windows start aggressively paging out virtual memory, unless you disabled paging altogether.
I believe it happened because the logic for virtual-memory handling was intertwined with the logic for implementing PAE. Prior to ~Win8, there were two operating modes: one for low-memory systems whose history went all the way back to Windows95 (and allowed it to limp with a mere 4mb of ram), and one for high-memory systems where PAE was assumed to be used (even though Microsoft ultimately disabled PAE in most systems).
When 64-bit Windows arrived, Microsoft started with the "high-memory" code, ripped out the PAE-implementation logic, but ran out of time/developers and left the PAE-imposed LIMITS in place. It wasn't until ~2014 that Microsoft finally got around to refactoring it to remove those limits. And even now, many of those limits still exist by default unless you disable them via registry keys because some popular older apps crash if they're disabled.
Don't believe me? Do a virgin-install of 64-bit Windows 10 on a new PC with 32gb of RAM, then TRY to get Windows to physically use more than 4 gigs of it without swapping (and without registry-hacking or disabling virtual memory altogether). What Windows NOW does by default is to swap pages to UPPER ram, with slightly better logic to decide when it needs to swap those "paged pages" to the ssd/hard drive. I think it also now can do most of it via register-manipulation instead of copying. But at the end of the day, Windows STILL defaults to treating your glorious 32gb ocean of ram as a glorified ramdisk unless you force it to do otherwise.
In case anybody's wondering why the mode is disabled for laptops, it's because quite a few literally *can't* run for extended periods of time at full-bore speed without overheating. For YEARS, companies like Toshiba were advertising specs based on CPU specs, but underclocking them behind the scenes to reduce heat. Most current laptops can (mostly) run at full-speed without crashing, but will gradually cook their electrolytic capacitors & other components to death if you insist.
Now, the big question is... what's Microsoft's definition of "laptop", and how does Windows decide? Is a mini-ITX "desktop" with mobile i5 a desktop or a laptop? How about a Clevo-made gaming beast with a "desktop" i7 & discrete (but semi-proprietary) graphics card that -- electronically -- is still PCIe (with nonstandard form factor)?
I wouldn't mind paying someone like Amazon or Google $10/month for access to every meaningful newspaper in America (with Google dividing it up among the papers I read that month), but I refuse to get sucked into a half-dozen monthly subscriptions... especially when seemingly all of them are "pay {some reasonable} rate for the first {n} weeks, then {get ass-raped} thereafter until you notice and cancel". I MIGHT do it if there were an option to automatically end the subscription once the promo rate expires, but over the past few years, I've gotten to the point where I automatically tell anyone trying to get me to sign up for teaser rates that silently go up to just go fuck themselves and die. I fell for subscription scams like that all the time when I was younger, but now it just seems like total bullshit and I refuse to put up with it anymore.
You can blame Verizon for plenty of shitty things, but CDMA2000 isn't one of them.
Verizon Wireless began its life outside the NY metro region (and probably WITHIN it) as "PrimeCo" -- a company that was "CDMA" from day one. At the time PrimeCo launched in the mid-90s, it had four real-world alternatives:
* Analog. Never a real alternative, but listed for comprehensiveness.
* 1G GSM. Nearly-irrelevant outside of Europe. Almost by definition, mid-90s 1G GSM mandated frequencies that would have been impossible to license in North America. Even IF a US carrier used "GSM" with "North American" frequencies, the likelihood that a non-North-American "GSM" phone would have been interoperable was approximately "none". Compatible dual-band phones eventually appeared, but were the overwhelming exception rather than the rule until well into the early 2000s.
* TDMA. In the US, used primarily by AT&T. Basically, AT&T skimmed off the low-level parts of 1G-GSM, lopped off the parts that were needed to permit interoperability with other networks, and tweaked the standard in ways that allowed them to exceed the 35km max tower-phone distance limit that was baked into GSM.
* IDEN. Was basically another semi-proprietary flavor of TDMA whose main selling point was its ability to handle point-to-point communications among employees all connected to the same tower without incurring airtime charges. Popular with businesses, but even in its heyday, nobody expected it to have a long-term future in the US.
Compared to CDMA, both 1G-GSM and TDMA were archaic & primitive. CDMA is SO advanced (compared to TDMA, which GSM and IDEN were both flavors of), it almost looks like pure black magic:
* Soft hand-offs -- switch from tower to tower without anybody noticing. With GSM & TDMA, you heard a distinct click as you were handed off.
* Few hard constraints about tower placement and spectrum allocation. With GSM & TDMA, tower location & spectrum allocation has to be carefully planned, and reconfiguring later is a Very Big Deal. With CDMA, you can literally solve congestion problems by semi-randomly throwing new towers into a congested area & the network will literally "fix itself" (that's not to say good planning is irrelevant... good planning gets better and more cost-effective results... but with GSM/TDMA, it's absolutely 100% mandatory and non-negotiably REQUIRED).
* More efficient spectrum reuse. Simply put, you can transmit more total bits among all active users over the span of a second using a finite chunk of spectrum using CDMA than TDMA/GSM/IDEN.
The bigger clusterfuck in the US happened due to spectrum chaos and Qualcomm's certification policies. In theory, almost any top-shelf Android device sold since ~2013 is technically capable of doing 3G GSM and CDMA2000-EVDO on any network in America... except they aren't. Sprint's network won't allow customers to "natively" use any phone not literally sold by Sprint (eg, Sprint will allow one of the few remaining Canadian Telus CDMA2000 phones to roam on Sprint, but if the owner moves to the US, Sprint won't allow them to sign up for service using the phone that roamed perfectly well on the same network). Likewise, Verizon will grudgingly allow you to use an unlocked CDMA2000 phone on their network... but they won't lift a finger to give you the SLIGHTEST bit of help, and a non-Verizon CDMA2000 phone will NEVER be able to use EVDO on Verizon due to the way they implemented authentication (only CDMA2000-1xRTT, which is about 150kbps). An unlocked AT&T phone might be WILLING to do 3.5G HSPA on T-Mobile with a T-mobile SIM card, but only on frequencies used (or formerly used) by AT&T. An unlocked T-Mobile[US] phone might be willing and able to do 3.5G HSPA on AT&T with an AT&T SIM card, but nevertheless be unable to do LTE at all REGARDLESS of frequency due to Qualcomm's licensing policies.
Qualcomm's licensing policies deserve special note. Basically, Qualcomm licenses radio modem firmware to CARRI
Musk's brilliance was having the Falcon's rockets designed for vertical soft landings. Among other things, this makes a certain degree of failure-risk during launch more acceptable.
With conventional rockets, rocket failure (of almost any kind) during launch almost inevitably meant payload loss. If you were *lucky*, you might be able to make a last-ditch effort to get your satellite into some kind of random orbit that might have value to a buyer SOMEWHERE, as opposed to writing it off as a total loss from day one.
With Falcon, if something goes wrong during launch & SpaceX realizes that they won't be able to get a satellite payload into its intended orbit, they could conceivably abort the launch, bring the whole thing back to Earth for a soft landing, and make another attempt to launch the same satellite another day. That's HUGE.
If nothing else, it means there's now an incentive to bring expensive satellites from failed launches back to earth, instead of making a last-ditch attempt to launch them into some random orbit anyway on the slim chance they'll end up being worth something to someone & not just end up as more space junk. It also means many customers will be able to roll the dice and skip buying a second satellite as a launch-spare, since "launch failure" won't necessarily mean "satellite loss".
In effect, SpaceX has potentially reduced costs at BOTH ends... rockets and launches themselves are cheaper, and the consequences of a "softly" failed launch are potentially reduced as well.
Bitcoin's fundamental problem is that if it doesn't crash & lose its value anyway, it'll eventually become too valuable to use as actual currency.
Bitcoin's proponents point to its decentralized nature and lack of government control, but overlook the fact that its blockchain can't scale. It's ALREADY staggeringly huge (something like 200 gigabytes the last time I checked). Sure, you can delegate responsibility for checking it to some other device (or third party) instead of hauling the whole thing around on your phone, but THEN you become dependent upon your ability to trust that third party AND its ability to keep bad guys from hacking them (to show that a spent bitcoin is still available, long after it has ceased to be). The second part is what fucks you if you try using YOUR OWN off-device infrastructure (like a server at home) or run by some public-spirited organization (that nevertheless assumes zero liability for anything that Goes Wrong)... not even large banks that actively TRY to prevent hacking attacks are able to succeed with only technical defenses. The entire credit card industry works because they have enough capital to absorb losses & use the legal systems of their respective countries to keep crime down to a tolerable slow burn. Your personal long-term ability to keep your own server hackproof (while nevertheless keeping it accessible to you when you're away from home) is approximately zero... eventually, someone will find a way to compromise your phone, your server, the PKI it depends upon, and/or anything in between... unless (or quite probably, EVEN IF) you dedicate your life to maintaining its security and integrity.
And in the meantime, you're going to spend more and more time waiting for transactions to clear unless you pay someone to expedite the transaction and assume at least some risk.
So, we get to problem #2... Bitcoin transactions can't be added to the blockchain until someone else manages to mine a new Bitcoin. As time goes on, this will become exponentially more expensive and difficult, so the fees they'll charge will go up & overwhelmingly impact otherwise-small transactions. If the cost of completing a transaction approaches something like 0.1 Bitcoin, it'll cease to be cost-effective to trade anything less than several Bitcoins at once. At best, Bitcoin becomes a virtual backing currency that only large traders can afford to exchange. To an extent, this has already happened with companies that use commercial wallets to handle microtransactions. And the companies running those microtransaction wallets are just as vulnerable as everyone else, except they're even BIGGER targets and even MORE hopelessly outgunned (think: homeowner vs armed burglar, compared to small business vs organized crime shakedown).
Long-term, Bitcoin will probably survive and pivot to new primary users when others fall, but it's hopelessly naive to think Bitcoin will EVER be usable for casual purchases. Commercial services that absorb risk to expedite transactions will end up costing as much as Visa or Mastercard, and be every bit as regulated by governments. As an almost-free guerrilla alternative to Paypal, Bitcoin is a long-term failure by design. As a virtual reserve currency, it might survive... but whether Bitcoin will ever become more cost-effective for businesses and banks to use for real trade than US Dollars, Euros, or any other currency is far from certain.
> (I'm not sure why one would want pages in a text heavy book).
Historically, technical books follow the pattern of, "Display a diagram, chart, or photo on one page, and talk about it on the facing page". Reading such books is MUCH harder when you can only see one page at a time, and trying to do it on a device that takes hundreds of milliseconds to flip between pages will slowly drive you mad. Slow page-flipping times is the #1 reason why reading technical books in ebook form is such an awful experience.
For an example of how fast eBook readers NEED to be able to render in order to not completely suck, check out this video: https://www.youtube.com/watch?...
I'm pretty sure the guy who made the video pre-rendered the entire book to bitmap images, because I doubt whether ANY tablet available in 2012 could have rendered pages from a typical PDF textbook quickly enough to render every page from scratch in realtime.
Was NASA *seriously* planning to launch a shuttle per week using the fleet of shuttles we had, or were they planning to build a much, MUCH larger fleet?
Considering how much large-scale refurbishment/remanufacturing the shuttle itself ended up needing after every launch (esp. the heat shield tiles in later years), I can't see how they could have even FANTASIZED about sustaining that kind of launch schedule with such a small fleet.
Did they strip out the Tesla's main battery (to reduce weight), or is that what's powering the transmitter & camera? It seems like a fully-charged Tesla battery capable of driving 200+ miles at 80mph SHOULD be more than capable of powering a camera and 100-watt transmitter for quite a while (100 watts doesn't sound like a lot, but it's actually the same amount of power used by the Mars Reconnaissance Orbiter when it uploads image data from the rovers back to NASA via the Deep Space Network).
For ebook-reading, 4:3 is actually a superior aspect ratio for most content. If you're scaling to fit 1 page on the screen, a 9" 4:3 is like a 10-11" 16:9.
For ebook-reading TWO pages side by side, a landscape 3:2 aspect ratio is ideal, because it's like two 4:3 (3:4) portrait displays side by side.
In any case, high-dpi (200ppi or better) is important... reading text on a 18" 1920x1080 display is like trying to read 1024x768 text on a shit 15" Packard-Bell CRT with .39mm dot pitch circa 1998. For all intents and purposes, a 240ppi OLED is more crisp for text than a 300dpi laser printer with premium paper (because it's 240ppi 8-16 bit grayscale, vs 300dpi 1-bit black/white).
The main problem with tablet-as-ebook-reader: they need bigger high-dpi screens (13-14" would be nice) AND fast primary storage & cpu. A tablet as aa ebook reader might spend 99% of its time doing nothing... but when you go to flip pages, it needs 10-20ms render time, MAX. Longer might be ok for things read serially (like fiction), but current ebook solutions SUCK for random-access reading, like technical manuals.
The Chuwi Hi12 has a good 12.1" 2160x1440 display at a fairly sane price... but it's an older model, and its other specs aren't much to write home about (mediocre wifi, average battery life, and slow microUSB charging via a port that doesn't inspire much confidence in its likely longevity.
That said, it's decent for 2-up ebook reading (a bit on the heavy side, but not unbearable).
No, YOU'RE the one who's full of shit. Lego Legends of Chima isn't playable on a factory-fresh 3DS until you allow the 3DS to fetch some mandatory update from Nintendo. It doesn't require continuous online connectivity to play it, but until the update is done, you won't be allowed to play the game. If you had a new-old-stock shrinkwrapped 3DS that spent 30 years in a closet and tried to play that game for the first time sometime around the mid 2050s, you wouldn't be allowed to do it.
As far as I can tell, Nintendo has an incrementing counter for patch level, and 3DS cartridges can specify a minimum patch level. If you put the cartridge in a 3DS that hasn't been patched to that level, the game won't run until the 3DS finds some way to connect to the internet, download the patch, and install it. Thus, if you're attempting to play a game for the first time that was released late in the 3DS's official life and requires a patch level that's higher than what that new-old-stock 3DS shipped with from the factory (or even a non-shrinkwrapped old 3DS that never managed to get updated to a newer patch level), you aren't going to be able to EVER play that game without some major hacking. Given Nintendo's past behavior, the likelihood that they'll have a server up and running to handle 3DS patch downloads circa 2050 is somewhere between "slim" and "none"... as is the likelihood that Nintendo (or whomever owns their copyrights at that point) would tolerate allowing anyone to independently distribute their old firmware updates after they themselves have discontinued support for them.
So fuck you. If you think this is "fake news", you're probably dumb enough to BE a prime target audience member FOR "fake news".
In the past, consoles either had no system software or shared libraries at all, or publishers were expected to distribute their games along with any patches to the baseline system software and libraries that their games depended upon. It wasn't until consoles gained network connectivity "out of the box" that companies felt entitled to start making them dependent upon post-manufacturing updates that could be made unavailable at any time without recourse from consumers.
IMHO, the manufacturer of any device that's unfit for its advertised purpose without additional downloads should be required by law to provide SOME "reasonable" way for a knowledgeable (not necessarily technically-clueless) end user to obtain and apply those updates on his own, LONG after the vendor itself has ceased to provide them online. For example, suppose a scenario like...
1. Hypothetical IETF RFC for "Post-EOL Firmware Update Service" that describes a protocol for a service you can run on a TCP/IP network and serve a firmware update to a bootloader long after a device has ceased to be supported by the original manufacturer.
2. A vendor like Nintendo escrows its firmware updates with the Library of Congress, and pays some nominal fee like $100 to register them. Ditto, for any software vendor who ships software that can't actually run without some mandatory update that isn't shipped along with it. Every 30-90 days, the LoC publishes the escrowed code and sells it to distributors who can resell it, give it away, or otherwise do whatever they want to with it. The LoC also maintains an index of escrowed code.
3. Years later, someone buys an antique 3DS at a flea market, along with some old games, then discovers he needs a patch to allow them to work. After doing some research, he learns that he needs to update that 3DS to patch #24.8.17.9 or newer. He could technically order it from the Library of Congress for $850, but he can instead get it free from "ancientfirmwareupdates.com" in return for watching 15 minutes of ads (because companies would buy and archive that published data, then redistribute it themselves with the LoC's full blessing). He sets up a server that implements the required service, points it to the patch data, and sets up an equally-antique 802.11n access point with SSID "3DSUPDATE". After consulting
And for approximately $25, you can buy a USB3.0 BD-ROM + DVD+/-RW drive the approximate size of a CD jewelbox that can also do CD-R(W).
You don't buy CDs to listen to directly... you buy them to get a clean source that can't be arbitrarily taken away from you {n} years from now because some company decided that "for life" means "the life of our product, as we define it" (e.g., Zune), shuts down their DRM servers, and leaves you with either nothing at all, or (at best) one final, fragile copy that'll be gone forever when the shit electrolytic capacitors (or glued-in battery) dies 2 or 3 years later.
This past Christmas, I spent a day playing with my old C64 & Vic-20. Both worked perfectly, and so did my old 1702 monitor. My old Odyssey2 (Videopac, in Europe) and RCA Studio II worked, too. It was a sobering experience when it sank in that there's probably not a single goddamn computer or game console you could buy today, put (shrinkwrapped) in a closet for ~35 years, and have ANY reasonable hope that it will actually WORK (and be usable "for real") when you power it up for the first time. Even a goddamn Nintendo 3DS refuses to let you do anything with most games until it connects to the internet & updates... once Nintendo shuts the servers down someday (like they're doing with the Wii's e-store this year), bye-bye system. Try doing anything useful with a Logitech Revue, a Zune, or a "WMV-HD" disc from ~10 years ago (Microsoft shut down their DRM servers & said 'fuck you' to customers, so they're now unplayable). This is why I'll NEVER spend money on full-priced content tied to Microsoft devices or services again, and why Google will probably never convince me to take their future media devices seriously. I'll throw down $7.99 for a used game, maybe, but never more than $29.99 (very rarely, more than $19.99), because I now assume anything I buy will be taken away 3-5 years from now & devalue it accordingly.
I can think of one area where reduced costs of long-haul trucking could result in exponentially greater utilization by consumers: storage. Especially if electric trucks had enough battery capacity to recharge in places where electricity is cheaper than nearby areas. Right now, we already HAVE "cube" storage... but the cubes themselves are still stored somewhere semi-nearby. If the cost of trucking those cubes 250 miles were only slightly greater than the cost of trucking those cubes 20 miles, instead of storing those cubes in warehouses that are only slightly cheaper than the area where they came from (say, New Jersey vs Manhattan), you can send them to a vast, sprawling, robotic storage facility out in the middle of rural Pennsylvania. It's kind of a contrived example, but shows how if something becomes cheap and easy enough, brute-force solutions that would be cost-prohibitive NOW can become financially viable.
The CODEC is another potential snag. Lossy algorithms like MP3 make a POINT of mangling and filtering away things that are "inaudible", especially if there are lots of OTHER things going on in the audio at that instant.
I'm not sure about AC3 & Dolby Digital, but I know that MP3 absolutely DESTROYS phase relationships used by algorithms like Qsound (for a very audible example, find a CD of Madonna: The Immaculate Collection, rip it, compress it to MP3 at 384kbps, 256kbps, 128kbps, and 64kbps (all CBR, not VBR), then compare them all to the original.
In the uncompressed CD audio, it feels like you can reach out and touch the audio.
At 384kbps, the spatiality is gone, but the stereo separation and frequency response are basically intact.
At 256kbps, the stereo separation "comes and goes"... mostly intact, but occasionally flattening out.
At 128kbps, the stereo separation is practically GONE, and you can hear the high-frequency response falling away.
At 64kbps, you might as well be listening to a cheap 1970s transistor radio (FM) with a 3" speaker.
I'd expect that the effect might actually be WORSE with bitrate-starved Dolby Digital Surround and DDS, because having multiple discrete surround channels partially hides the effect of mangling stereo phase relationships. The compression algorithm might tolerate and preserve higher-frequency audio on certain channels, but not others (e.g., if you mixed a 19KHz sine wave into the center channel or surround channels, it might be attenuated away, but might be at least partially preserved if you mixed it into the front left or right channels).
This is part of the reason why Dolby ProLogic Surround (note1) works somewhat well with audio from high-bitrate sources, but falls flat on its face & dies with many compressed-audio sources. For example, try watching a Hollywood Movie streamed via SlingTV with your AVR set to decode it as ProLogic surround sound. On many channels, the audio will just completely fall apart, generating endless pops and other audio artifacts on the surround channels unless you forcibly disable ProLogic and make the AVR treat it like plain stereo (most AVRs will default to treating 2-channel PCM [streamed as AC3, decoded to PCM2.0, and sent over HDMI by a Roku] as de-facto ProLogic, because often the 2-channel stereo track IS surround-encoded-as-ProLogic. SlingTV's audio bitrate is too low, and mangles it into oblivion.)
TL/DR: high-frequency audio is one of the first things that gets kicked to the curb (or mangled) by bitrate-starved surround-sound compression algorithms.
-----
note 1: Back in the late 1970s, Dolby came up with "Dolby Surround" for theaters. Broadly speaking, it allows you to encode a low-fidelity third (rear surround) channel into the left and right stereo signals in a way that makes it sound like reverb if played back by stereo gear. The catch is, that rear surround channel can't deviate from the audio from the left and right channels by more than a few dB, or it will audibly "leak" into the front speakers.
ProLogic takes the "Dolby Surround" signal, decodes the third channel, then runs everything through a DSP to further transform it into something that (vaguely) resembles 5.1 surround. It's a pale imitation, but when done well, can sound better than plain stereo.
Because analog Dolby Surround (and ProLogic surround) is mastered into the analog stereo soundtrack of a movie, a lot of movies distributed on LaserDisc and VHS ended up having passable surround sound by default if you ran the stereo audio through a surround sound processor. It was basically a lucky accident, but it worked.
DVDs arrived, and digital surround (primarily Dolby Digital 5.1 and DDS 5.1) came along with it, and ProLogic became something that was "there" by default in the stereo soundtrack (mostly, because once you mixed it for 5.1, downmixing it to ProLogic was something any decent software could do automatically). Blu Ray took this and extended it further to 7.1-channel DD+ and
Exactly. Long before ANY colonists arrive, the terrain surrounding that first colony will be LITTERED with hundreds of earlier landing craft that delivered robots to begin building & running its infrastructure. There will be THOUSANDS of virtual colonists who've spent YEARS vicariously "living" on Mars by overseeing those robots remotely from Earth (obviously, not in realtime, but spending hours per day reviewing "their" robot's past Sol). By that point, there will even be three sources of recyclable resources:
1) a distant area where the earth-mars rockets used to transport skycrane-deliverable robots were crashed.
2) a nearby area where the descent vehicles were crashed after delivering robots in a skycrane-like manner, and where cargo craft landed normally.
3) All the robots that quit working after arrival.
Sending ANY humans before the robots have built suitable living & working quarters & have things like life support & power generation working flawlessly would be stupid & irresponsible.
You're assuming that whomever started a global thermonuclear war (or their target) wouldn't have a missile or two aimed at Mars, just to take out the enemy's colonies, too.
Imagine it's 1980 in a parallel universe where only the Soviet Union had a colony on Mars. Do you really think the US military wouldn't have classified it as a major target & had at least 1 robotic rocket in its arsenal capable of flying to Mars & launching one or more missiles at their colonies from space? It's sad, but it's totally naive & unrealistic to think otherwise.