or you buy something, but what they DONT tell you is the product you purchased is located half way around the world in china, but you dont know that until they send you the product shipped notification and you see it just left the facility in china somewhere, and you dont get the product for 3 months,
i quit buying from amazon because of their dirty shenanigans, to hell with bezos the bozo i dont need his clown tactics when i want to buy something
Perhaps it pays to see who you're actually buying from then. Because Amazon products are shipped from Amazon's warehouses. If you buy from a third party, then all bets are off.
And no, there's no link on Amazon that is not subject to this - if Amazon runs out of a product, they will quietly substitute in a 3rd party seller Or if a third party seller has it cheaper, Amazon might prefer them. I've had to go to the "New and Used offers" link more than a few times to actually buy from Amazon.
Of course, if you don't have prime, don't regular Amazon shipments take 3 months anyways?
Actually, I should also say this is why most tech workers believe they are underpaid as they know of people in silicon valley earning twice or more their salary.
I think if you ask anyone, they'd think they're being underpaid. Not just tech people, but anyone. From the small business owner who barely makes minimum wage (running a business is hard work), to the janitors who break their backs nightly mopping floors to the CEOs who always believe they need more.
I don't think there's anyone who would answer that they make enough money right now.
I'm beginning to think that we're reaching the limits of what we can do with the laser lithography method of silicon IC creation. For instance look at the problems Intel is having with 10nm fabrication right now. Perhaps the way forward is straight out of science fiction: a matter compiler/3D printer-like approach, where an integrated circuit is built up an atom or a molecule at a time? Pure imagination on my part, but is it really out of our reach?
Actually, modern ICs are done at the atom level buildups - to create the strained silicon that we use today requires subjecting the wafers to vapor deposition that literally grows the surfaces atom by atom.
The problem with 10nm is not the photolithographic process - we've outdone ourselves in this technology. It's the fact that at 10nm, transistor gate lengths are starting to be measured in the realm of atoms - with gate insulation thicknesses being a handful of atoms (which can easily lead to leakages). Add to this quantum effects - tunneling and the like and you get leaky transistors that consume a lot of power and generate heat.
Also remember that the only circuit elements that use the smallest dimensions are transistor dense ones - memory blocks. Things like general logic in a processor a wire-constrained (you are limited by how much wiring you can stuff between transistors and not the transistor size themselves), which is why they always fab tons more transistors than are needed, but still way larger than minimum size.
Thus, leaky transistors are in places where you don't want them to leak - generally memory cells where leaky charges mean bit flips.
Brandname matters. Intel/Nvidia is the best gaming combo. It just works and games are tested and optimized for both as they own 90% of the CPU/GPU market. Corporations buy whatever HP and Dell throw at them. THey like their Intel contracts and want to stay good with Intel for cheap pricing.
Intel means reliability to corporate buyers. It works well and everyone else uses it so they need to use it too. Brand name again and last drivers and issues are less with Nvidia and Intel. Always have. AMD is playing catch up but when you buy Intel or Nvidia the drivers are optimized on day one historically.
AMD for these reasons are a tiny tiny player
The other thing is Intel can deliver. As in, if the OEM buys a million CPUs, Intel will deliver a million CPUs.
Intel has plenty of fab capacity. AMD is fab constrained (and always has been).
It's one reason why Apple went with Intel and not AMD - Apple has traditionally been at the mercy of CPU manufacturers (first Motorola, then IBM), some famous shortages of high-end machines were common in the pre-Intel days.
No doubt Dell and HP and others have AMD offerings, but I'm sure their volumes are much lower and generally only concentrate on the midrange models, the ones that AMD can likely provide in bulk.
Yes, there's a lot of fanboy and underdog going on, but if HP/Dell/etc decided to offer high end machines, AMD would actually get into serious trouble - because they'd run out of the high end chips, and be stuck with a bunch of midrange and low end chips they can do anything with, and unhappy customers all around. Enthusiasts would balk at paying $$$ for high end chips (think the old graphics card markups were bad during the cryptocurrency days?), and a ton of low end chips would flood the market no one wanted. (You could easily get 1050s and lower GPUs for regular price...)
OpenGL is shit and has always been shit. It has a convoluted API that was designed around configuring fixed-function 3D rasterisers and is ill-suited to programmable GPUs, multi-threading and has extremely high overhead for modern graphics pipelines from the multitude of function calls and GPU state validations.
Metal, Vulkan, D3D12 are all very similar in the way they offer lightweight APIs that are a minimal abstraction of modern programmable GPUs and are designed to work in multi-threaded, multi-process environments.
Correct. OpenGL is an antique API - if you have a modern high end video card (like a GTX 1050 or higher) OpenGL will run like crap on it - it just has too much overhead causing most of the power to go underutilized.
The big problem is what OS X is going to use - Metal, Vulkan and the like all came out around the same time because of the issues of OpenGL One should note when Metal came out, Vulkan was actually AMD's API set - it was donated to Khronos to offer a standardized next-generation API set, and renamed to Vulkan. We are in a huge transition period where legacy apps will need to be ported over to take advantage of modern video card performance.
One millennial options trader lost $180,000 overnight
That bit really made my day.
Let's see, he lost $180K, and FB went down 19% that one day, which would imply just under a million dollars or so in options trading.
Sorry, when you're dealing with those numbers ( and it's options, not real stock), it's somewhat hard to feel sorry anymore. Especially since if you read up on the guy and his celebrations of $300K in gains. $180K is a lot, but if you're doing that level of OPTIONS trading, you're probably fairly well taken care of and that $180k might just mean you skip today's Bentley purchase. (The people who do those sort of trades are either institutional, or already multimillionaire rich).
It's like people who complain a tax increase will cost them $50,000 more (happened). Turns out if you do that calculation, the guy was already doing about $2M in income so while the numbers are impressive, the real meaning behind them is hidden.
And yes, it always helps to do that calculation. Next time someone claims big numbers, find out what that tax increase actually means and you'll probably find out they're not going to be hurting quite so much.
We're just paying a 3% fee to a 3rd party to make the payments for us.
There is an easy fix for this. Require the end user to pay the 3% fee. A few places do it but I believe it's against the rules of most major credit card companies. If more places gave a discount for cash, a lot more people would pay with cash. As it stands now, with credit card rewards, there is effectively a 1% or so discount for using a credit card.
First of all, cash isn't free to handle. If you're a small business, it might be close to free, but only because you don't have much of it. When you're handling a few thousands of dollars of cash daily, things get MUCH more interesting.
And it's easy to tell who's the "free" department - you can tell if a business simply turns off the register at the end of the day, or if they actually count the takes. If they don't count the takes, then in general they don't make enough cash (because it's not worth counting).
But once you do start raking in cash, you start to have to count it, which means you need to train employees how to handle cash. You need to tell them how to set up the register, how to ensure their takes are accurate, and how to sign off the register. This again is easy to see - when they run out of coins you note they always take a 410 roll of coins, but at the same time, they give the donator a $10 bill to equalize the transaction.(because every dollar has to be accounted for). And yes, when the register and cashbox do not agree, it's a very serious offense, minus some small percentage error (because people do realize short-changes and mistakes happen, but it shouldn't be more than a couple of bucks or so).
But you also have the problem - you have a few thousand dollars in cash now, and you have to deposit it at the bank. So someone has to go to the bank, wait in line, and make the deposit (which has to be counted, verified, etc). So that's someone's hour out of their day going to the bank, parking, lining up, etc., and hopefully not getting mugged or robbed along the way (another cost of handling cash). Or if you're lucky, you've got night depository permission so you can use those outside deposit boxes. But again, prime spot for being mugged (and it's after hours, too).
Hey, if you make enough money, perhaps you can call an armored car to take your cash for you to the bank, but that too is another cost for handling cash.
Oh, and let's not forget the whole robbery aspect - always a problem with cash-only businesses.
And finally, there's the problem of counterfeit bills. You might not think it, but large bills might be common - if you buy $50 worth of stuff, you would expect to get paid with $20 bills, $50 bills, and $100 bills. But $100 bills could be counterfeit, yet another cost (because counterfeit bills are non-redeemable, and it could've been changed with legitimate bills). Sure you can try to avoid taking in $50 or $100 bills, but if someone comes to you with $75 worth of purchase, they may not have $75 in smaller bills.
Now, the laws changed here recently (Canada) where retailers can charge the extra 3%, but it turns out most don't. Not because they can't, but because they realize that most of their transactions are credit and debit (which incur a fixed fee at least), and they fear losing that business. (And yes, businesses that were formerly cash-only have gone into accepting debt and later on, credit cards did see an uptick in business - both in number of customers through, and increases in size of transactions).
And to be honest, the ones that do charge, I tended to shop there less - about the ones that did were computer stores and they always did "cash discounted" pricing. It was always annoying to have to plan a visit to the bank to do a cash withdrawal in order to shop there, and they eventually lost me out to online shopping where the credit prices were included AND I had to pay shipping. (Not having to find parking in some obscure neighbourhood was a bonus, too).
They failed to meet their deadlines and other commitments before the union fight started.
Chances are, Charter was probably going to slip under the radar, then they started the union fight. The union reacted by filing a complaint with the PSC which led to an investigation and all the charges.
Charter was probably hoping to skirt by doing crap and having no one noticed, then they started fighting with the union and someone there realized Charter isn't quite on the up and up.
Usually it's just a pressure tactic, but I guess this time it backfired on Charter
My guess is that the theaters imagined it would mean more revenue in the form of snacks, since ticket sales are already eaten by the cost of getting a copy of the film. Then again, if someone is cheap enough to use the app rather than buy movie tickets, they're also probably not buying snacks/probably sneaking them in.
That's only if MoviePass worked with theatres. Surprisingly, no theatre chain is working with them, and AMC blocked acceptance of MoviePass completely.
The thing is, MoviePass just has the wrong business model, period. Ticket prices, at least in the first two weeks, generally all go towards the studios. There is no way the theatre can offer MoviePass a break on ticket prices because of it.
The only time they can takes place on the third week or later, and that's when the theatre starts getting a small cut of the ticket revenue (generally to make up for smaller crowds and lower concession revenue).
It's generally fine - concessions bought during the first couple of weeks makes up for the loss in ticket revenue which is why theatres don't really mind giving 100% of ticket sales to the movie studios.
But even if there's an empty seat, as long as it's empty, the theatre pays $0. If that seat gets occupied, even if by someone with a discount ticket, the theatre must pay the full ticket price. So even a flop of a movie that doesn't sell in the second week, the theatre will rather show to an empty theatre (and pay $0) than have people come in.
This isn't an airline or transport business where an empty seat is a loss of potential revenue. An unsold ticket means an empty seat and it means the theatre does not owe the studio that money.
For MoviePass to survive, they must block the ability to view a movie in at least the first 14 days (the days the theatres make no money off tickets, and would rather have an empty seat and pay $0 to the studios for it over a discounted ticket where they'd have to make up the difference). They should also realize that concessions are how theatres make money, so offering free small popcorn or something to encourage sales of concessions would be necessary. (A small popcorn would lead to at least a drink sale at full price, and if people are going with friends, generally more concession sales as friends would want to buy stuff as well).
As for data mining, I'm sure the theatres have good ideas what their patrons do. "Dinner and a movie" is a common phrase for an activity, which is why they often have partners with restaurants. And after a movie is over, if it's a late showing, generally speaking most patrons are either going to go home, or to a bar. But in general, I don't think it's too valuable because too many people are doing too many different things.
Slack is the cool player in the team chat space. In Digital-land everything is Agile and Devops and modular and distributed, so open team chat is the best method to collaborate rather than email. These new team chat platforms are not just apps like IRC, Skype, Messenger etc, they have all sorts of useful new features such and integrations, APIs, Bots, wiki, document storage, search etc.
And ensuring you have the latest and greatest hardware. I've never seen a chat client consume 30% CPU resources before, for what is effectively webchat. Other web chat things like Discord and such, no problem, they take less than 1% idle.
So if you're using Slack on anything lower than an i7 (i9 preferred) with 32GB of RAM, or preferentially a whole separate machine, it's painful.
And yes, webchat - because the "app" is actually a node.js/electron thing that wraps their webchat into an independent app. Which is preferable, since some browsers would just clog up if you used the web interface. The app just ran its own instance of Chromium independently so it wouldn't bog down your web browser.
As someone who thinks 75% of cases where email is used would be better using something like usenet[1], do these things leave a trace?
Because if they don't, they're useless.
Monday. Marketing: Make it all purple! Tuesday: Your boss: Why did you make it purple? Marketing are throwing a fit.
[1] Yeah, first rule...
Well, there's plenty of logging, and it's a bit cloudy too, so even if you joined today, you can still see previous chat logs (don't know how far back, but if you just keep scrolling back, more is fetched). So even if your boss bamboozles the new guy on the team, one he logs in to the chat, he can always scroll up and see what your conversation was.
So from this discussion it's interesting that hyperthreading has been stuck on 2 threads per core for so long. Why not 16 today?
Because of diminishing returns. It's called "Multithreading" and not "Multiprocessing" because even though it shows up as a separate CPU core, it's shared resources could benefit from threads of the same process far more than disparate threads, especially with respect to cache utilization.
It comes about because a modern CPU is super-scalar. That is, there are multiple execution units - often 3 (or more) integer units doing simple operations (adds, subtracts, bitshifts, boolean operations, etc), a couple of "complex" integer operations (multiply, divide), several load/store units (at least two each for loading and storing, and sometimes the effective address calculation uses the simple integer ALUs), multiple floating point units - again 2-3 for simple operations, 1 for complex operations and other units. Thing is, one program instruction stream, even with multiple instructions in flight simultaneously may not actually use all the resources. It's fairly cheap to tack on another front end execution unit and run another set of instructions through the same execution units, with the goal to have more utilization of the idle execution units.
Thus, disparate workloads generally work better - if one is more integer versus floating point, you can often run both together without them colliding. But if you're using all integer, and even worse, both are using the same units, then it will not speedup.
The Chinese learnt from the ZTE case that yu should do as little as possible buisiness with American companies. NXP is now a Dutch company, but it would become a US company subject to their idiot sanctions. Trump tried to prevert sanctions against ZTE but congress didn't let him. So now the US gets repaid for the congress decision.
Good, because ZTE violated all sorts of regulations that were inposed on it.
The only reason Trump supported ZTE was China invested $500M into Trump hotels and granted Ivanka Trump a rare Chinese trademark. That's it. A little backroom dealing and let's forget that we asked you nicely to not resell those parts to places we asked you not to. And it's not like it was the first time, either.
If you ever wonder why Trump seems to care for Putin, again, remember Russia is one of the few places where Trump's business are actually succeeding, quite well in fact. Everywhere else they are doing terribly. All this from a President who has not divested himself of his businesses while in power.
Congress simply grew a backbone and enforced the order. Or shall we say, China did not scratch their backs obviously.
Some slick games and gizmos used Flash. HTML5/JS versions of equivalent still seem glitchy and browser-version-sensitive to me. Whether it will eventually settle, or some new trend/fad will break it worse is hard to say.
It depends. If the grame or application is programmed as a web game, it will likely be glitchy.
However, the new trend is WebAssembly, and you don't code in it directly, but use a cross-compiler. You write your code in C and it compiles down to a bunch of javascript (using a highly restricted subset that runs really fast) and your code runs. plugin-free.
The Internet Archive has plenty of play-in-browser DOSbox based games done like this. There are also in-browser versions of MUNT and other music players.
And the whole reason we have WebAssembly was someone at Mozilla wanted to have Unity be plugin-free (which created asm.js). So the future of web games is likely to be games that can be run natively or be run in a browser if a native runtime is not available.
So, that would be around 10pm for me. Guess I get it faster because I missed the rush of orders during peak time. Wish the place where I get sushi delivered.
Yes, this data is valuable because you can avoid the rush and get your pizza faster, hotter and fresher. During peak delivery periods, they would have to produce several orders at once for a single delivery run and thus where you lived determined how long the pizza is sitting in the delivery vehicle. Avoiding the peak periods means the kitchen makes the pizza and it's delivered straight to your door since you're the only order at the time.
I recently held a cyber-security discussion at a local retirement community. We went over a lot of ways to stay safe online (e.g. don't open file attachments, don't download programs from random sites, if the deal sounds too good to be true then it probably is). I also covered phone scams. My co-presenter had happened to get an IRS phone scam recently and still had it on her voicemail so we played it for them. The caller claimed to be from the IRS, said she owed money, and threatened my co-host with arrest if she didn't pay up. My co-host never paid and is still "evading the cops" (translation: living normally because no police are after her). We might have scared the residents a bit with the threats we revealed, but we reiterated that using the Internet was like going through a big city. There's lots of great stuff to see, but you just need to be careful.
I'm always curious - the tax scammers never identify who is in trouble. They always say "You're in big trouble with the tax man and the police are after you". Yet, most households in the US consist of multiple taxpayers - most families have both parents that work, and the kids may hold a job (summer or part time) as well.
Does it never occur to people that the scammers never identify the taxpayer involved? I admit once I nearly got fooled (it was a slick non-computerized speaker), but then I remembered the guy left a message without saying WHO was in trouble. If could be all of us, in which case they would've mentioned all of us, but they just said "you" which was highly ambiguous. Not like the tax man doesn't know how many taxpayers are at an address, and to actually identify which one they want to talk to.
Of course, I suppose I'm curious now what happens if they say that and I simply reply "I'm not eligible to pay taxes" and then ask who they want to speak to.
Finally, back on topic - you should mention the real legal process - the taxman always sends you a letter first, and legally, they cannot send the police after you, at least not without you completely knowing about it. As in there would be letters, there would be court service, court dates, there would be court sessions, judgements, etc. It will never be just a phone call. Justice is not swift, either - the process will take years. And even if at the end you're guilty, they generally still don't send the police after you.
Tax scammers use to only target the immigrant community (who were generally ignorant of how the legal system works, and in many countries outside North America, justice is often swift like that (which is what the scammers relied on). Here justice is slow and methodical and you get plenty of chances to defend yourself.
Sequence and filmstrip methods use the same markup as a static photo. How would the browser distinguish markup that represents a static JPEG, PNG, or still GIF from markup that represents an animated GIF, a container for an image sequence, or a filmstrip?
Easy, you only render the first frame and that's it. There were options to not play animated GIFs since browsers were first released, so all Firefox and other browsers need to do is simply modify the image renderer to only draw out the first frame. They can draw a simple play button to indicate that there are more images waiting to be drawn onto the page, but are currently blocked from doing so.
No one said you had to render out every frame of a multi-frame image. Just render the first, offer to play back other frames and leave it at that.
I used to enjoy them, my best was 16 minutes and screaming abuse at the end, but I got 4 in one day, so I blocked all incoming international calls. They dropped off for a while, but now they're spoofing local numbers. Sometimes I answer, sometimes I don't.
I've got some hindi swear/abuse phrases ready to go (stuff like "you're the result of a toilet cleaner fucking a goat"), but my best was asking the girl from "Windows technical department" what her mother would think of her activities. She went silent for about 10 seconds, then hung up.
Amateur. YouTube has a few people who can get scammers on the phone for at least an hour, approaching 4 and a half. Though to be honest, a good chunk of it is simply dead air - they ask the victim to go out and buy gift cards or something (which is good for a couple of hours of scammers on hold).
And scammers have an "emergency out" - if they get exposed, they immediately lock the machine and syskey it to ensure it is unbootable. Of course, it all happens in a virtual machine, so restoring the damage is trivial, but it's funny.
Some YouTube channels: Kitboga. He's had scammers insult him, make death threats and all sorts of fun.
This guy investigates scammers computers. His latest video involves a bank login scam where he manages to install a RAT on the scammer's computer. He watches as a scammer attempts to register for and log into some elderly guy's bank account (luckily, the bank actually sends something to customers when this happens, so the scammer not only had no chance. He also alerted the bank who locked the online account for fraud.
Yes, it's fun watching scammer's computers and call centers get infected. Perhaps they can call themselves for tech support.
I admit, I get a lot of those calls. It's amazing how they always use the most robotic of voices. Also, apparently retailers are alert to the scam as well - several times they've been stopped when they see people buying thousands of dollars worth of iTunes or other gift cards. Though, someone was so embarrassed they made up a whole story about being "arrested" and transported and caused the police to issue an alert.
CPU throttling is designed to address thermal issues in mobile devices. If the CPU is throttling so much it goes below the lower threshold and is STILL too hot then there is a hardware design issue. I'm extremely curious what software could be causing the CPU to be hot even when it is throttled.
Except, the CPU was not overheating. The throttling did not happen because of the CPU
The CPU is powered by a Voltage Regulator Module (power delivery module). It turns out it is this chip that is overheating - when the CPU is going full tilt, it's demanding 125W from the VRM. This causes the VRMs to heat up and when they get close to their maximum thermal limits, they send a signal to the motherboard telling it to throttle the CPUs so they draw less power so the VRMs can cool down.
Part of thermal tuning is to adjust the CPU boost speeds such that it can boost to full speed, then throttle down slightly as the VRM and CPU heat up to a new max steady-state condition where the heat generated can be dissipated.
This is in part due to a documentation error in Intel's docs regarding max thermal power dissipation values.
The good news is if you tweak the throttle settings properly, you can keep the regulators from overheating, but the CPU still performing. This is what Apple did - they optimized the settings so the VRMs will not overheat and force a sudden throttling of the system. Doing this gives you a good 20% speed boost over the old models.
The bad news is if this was caught earlier so Apple could heatsink the VRMs to the CPU like they do with the GPU, you could get up to another 10-20% in performance because you can run the boosts longer since the VRMs would heat up slower.
If you aren't playing games on your laptop, then perhaps you are using a development IDE or even Excel. There are far more applications than just games which can take advantage of an Intel HQ line processor or a decent video card. I would almost venture to say that anyone who can get away with an under-powered power saving CPU could probably do their work on a cheaper tablet.
The thing is, 99% of the CPU is idling even when running an IDE or Excel. So much so even a complex worksheet may take longer to open on a thin and light laptop, but in general, Excel or the IDE can crunch while I'm thinking about my next move.
It's important for the app to open fast and display what it knows. Recalculate in the background because by the time I navigate around to find what I want, the CPU is probably done with the calculating and macros and all that.
I love myself a beefy laptop, but while it's nice and powerful and runs everything I threw at it, it wasn't the best. Carrying it around was such a pain it was easier to just leave it at my desk. Traveling with it mean having to plan beforehand how I was going to carry it - its heft and weight made for real issues. And forget about using it enroute.
Today's modern processors mean it's Good Enough. It may not crank out the FPS as much, but eGPUs help, and it means when move about, I can carry my thin and light laptop with me - it slips in my bag and it's not that much heavier with the heft of other stuff. I can take it out and use it when I need to - granted, it's not got the graphical power so I can't play heavy games, but I can do plenty of other stuff with it. I can even code using an IDE with it. Getting decent battery life (something my laptop battlestation never had - the battery was more of a UPS than anything) is a bonus.
Granted, it's not for everybody - there are people who need to run 3D applications, Photoshop and other things that need raw power on the go and would need a huge heavy laptop. I just found while it was great, it was also more limiting. Turned out for me, I needed less a luggable, and more a portable.
I'm sort of surprised to see Qualcomm releasing a discrete part at this level of integration (rather than "here's an entire cell modem on module so you can 5G your widget with minimum regulatory hassle and without adding an RF witch doctor to your traffic light control company's team" or "here's the silicon and some design guide docs; engineering support for larger customers").
Qualcomm makes parts. They sell parts. They do not sell modules. You cannot go to Qualcomm and buy a complete cellular module.
Instead, companies like Sierra Wireless, U-Blox, Widcomm and tons of others take Qualcomm's chips and turn them into modules of all sorts of form factors. Qualcomm provides support to those companies, while end users of those modules either use provided software by the module vendor, or write their own based on the 3GPP specification documents.
There's clearly still a demand. And the used games market has dried up because piracy is much simpler.
The demand is obvious, given the Nintendo NES Classic re-release (after that disasterous launch in 2006). Sure you can probably build a cheaper one, but again, Nintendo's version has all the ROMs licensed properly.
(And Nintendo didn't see Sega and Atari screwed themselves when they sold their stuff to AT Games, which is why those retro consoles never sold well - they may say 1000 games on the box, but it's really about 10 official real Sega Genesis or 2600 games, and 990 "inspired" games by AT Games themselves, so it's really a box full of shovelware). Whereas the NES Classic (and SNES Classic) contain nothing but real games.
And the re-release is done quietly this time around so it's actually possible to get them. Launch day stock lasted beyond the first day because unless you knew, it wasn't advertised.
Any why are we so quick to condemn actions like this, but cheer on the likes of people who go after GPL violators? it's exactly the same thing A GPL violator is a pirate - they've committed copyright infringement. Distributing GPL binaries without source is copyright violation - if you didn't agree to the GPL, your rights are what the law grants you.
There are several stories of Qualcomm and Samsung trying to produce anti-apple propaganda about this, because really they have nothing compelling to say against the iPhone X.
Qualcomm is doing it because Apple does deliberately slow down the Qualcomm chipsets. Apple does it to minimize differences between phone models (there's not really "one" iPhone X, but a couple of them, and depending on the carrier, they get one or the other).
This is in part because of the Apple-Qualcomm lawsuit. Qualcomm obviously wants Apple to go all Qualcomm, but Apple doesn't want to be beholden to a single supplier - they've been down that road before (Motorola, IBM both failed to deliver the product Apple wanted on time, leading to many shortages in the past).
Samsung's doing it because their Galaxy S10 is coming out sooner or later, so they need to ramp up their advertising.
Of course, one wonders if the many Android bootlooping problems could be solved with Apple's "performance management" solution. I know it kicks in if the phone fails to boot because of the battery (i.e., what would've been a boot loop is then managed by the OS so instead of a dead rebooting phone, you have a working, but slower, phone). Some phones that bootloop were solved by simply reloading a kernel that disabled the high-power cores, a crude form of performance limiting, really.
Mr Gruber's scenario of hidden corruption just isn't a thing.
It can be.
On Windows, where the default is Quick Removal, what happens is Windows disables write caching internally and doesn't let WriteFile() or Close() return until all the data buffers have been committed down to the media. This means when an application writes, the calls block (or the async I/O object is not signalled) until Windows is satisfied that it has pushed every byte of data down to the media. For USB devices, this means it sets the "Force Media Access" bit which tells the device to actually commit the data to the media. But, Windows assumes when the USB device replies, it's done (assuming no write caching done hidden from Windows). Windows is correct that as far as it can tell, it's done its job - the data has been pushed to hardware.
But as anyone can tell you, Windows 98 had a fatal bug where new PCs would shut down so quickly that despite Windows flushing all the caches, it would still shut down faster than the disks would finish flushing their caches to disk.
Similarly, if you know how SSDs work, you may want to wait a couple of seconds after the write completes so the controller can flush out the updated tables to media. Not doing so risks the controller tables being corrupted and drive nonfunctional, or the file being corrupted because of incorrect indices.
Finally, a bad Windows app can cause extensive wear on removable media because of this - if you write a small buffer, that buffer will go to media. If it writes another small buffer, then it forces another write
And on Linux and macOS, it's somewhat essential, since those OSes still enable write-back caching. Yanking the drive before unmounting will result in corrupt files or a corrupt filesystem.
Make me wonder what the consensus is on the wet nature of water or being punched in the nose by a biker.
Well, back before email, you typically sent a nice letter - handwritten or typed up and it was generally appreciated because it showed you took effort on it - after all, typing (or writing) a physical letter takes time, finding and envelope and paying for a stamp. It's why letters still generally have an effect.
But in this modern day of email, things are not so clear. Emails takes just a few minutes to type up, cost practically nothing to send, and may give an air of insincerity because of it. Especially when the recipient may be bombarded by hundreds of emails a day, so your note may get lost in the mess.
So sometimes a study like this helps clear the air that those little thank you notes that take almost no time at all to send are still appreciated. Of course, they still have to be personalized, so sending a mass emailed thank you note to everyone at your wedding thanking them for their gifts... may still come off as insincere.
The whole point of higher education is that you are getting educated from a reliable source, and that the tuition you pay justifies it. If colleges are just going to tell you to read Wikipedia for four years then why bother going?
Education is moving away from the 19th century module of facts and figures, and into the 21st century of analysis. It's moving very slowly, but it's become obvious it's the way to go.
Facts and figures were appropriate in the 19th century because it was hard to distribute, search and acquire information, so one could be more "educated" simply by knowing more. It's why tests are still a 19th century instrument of factual recall.
But today, we have facts and figures at our fingertips. There's no need to memorize facts and figures anymore, because it's all so easily looked up in an instant. Rote memorization isn't as useful.
Instead, the 21st century education system needs to go one further - given everyone has access to all the facts and figures, one needs to be able to analyze, evaluate and process them. The current education system is not set up to do this - to be able to take facts and figures, analyze them and figure out what's "fake news," "exaggerated," "contains omissions," or other errors. (Talking about real fabricated news here, not "news that puts Donald Trump in a less than positive light" "fake news"0.
That's unfortunately where college education generally comes in - in general, that's where the skills necessary to function in today's world comes in .
Slashdot Mirror
Perhaps it pays to see who you're actually buying from then. Because Amazon products are shipped from Amazon's warehouses. If you buy from a third party, then all bets are off.
And no, there's no link on Amazon that is not subject to this - if Amazon runs out of a product, they will quietly substitute in a 3rd party seller Or if a third party seller has it cheaper, Amazon might prefer them. I've had to go to the "New and Used offers" link more than a few times to actually buy from Amazon.
Of course, if you don't have prime, don't regular Amazon shipments take 3 months anyways?
I think if you ask anyone, they'd think they're being underpaid. Not just tech people, but anyone. From the small business owner who barely makes minimum wage (running a business is hard work), to the janitors who break their backs nightly mopping floors to the CEOs who always believe they need more.
I don't think there's anyone who would answer that they make enough money right now.
Actually, modern ICs are done at the atom level buildups - to create the strained silicon that we use today requires subjecting the wafers to vapor deposition that literally grows the surfaces atom by atom.
The problem with 10nm is not the photolithographic process - we've outdone ourselves in this technology. It's the fact that at 10nm, transistor gate lengths are starting to be measured in the realm of atoms - with gate insulation thicknesses being a handful of atoms (which can easily lead to leakages). Add to this quantum effects - tunneling and the like and you get leaky transistors that consume a lot of power and generate heat.
Also remember that the only circuit elements that use the smallest dimensions are transistor dense ones - memory blocks. Things like general logic in a processor a wire-constrained (you are limited by how much wiring you can stuff between transistors and not the transistor size themselves), which is why they always fab tons more transistors than are needed, but still way larger than minimum size.
Thus, leaky transistors are in places where you don't want them to leak - generally memory cells where leaky charges mean bit flips.
The other thing is Intel can deliver. As in, if the OEM buys a million CPUs, Intel will deliver a million CPUs.
Intel has plenty of fab capacity. AMD is fab constrained (and always has been).
It's one reason why Apple went with Intel and not AMD - Apple has traditionally been at the mercy of CPU manufacturers (first Motorola, then IBM), some famous shortages of high-end machines were common in the pre-Intel days.
No doubt Dell and HP and others have AMD offerings, but I'm sure their volumes are much lower and generally only concentrate on the midrange models, the ones that AMD can likely provide in bulk.
Yes, there's a lot of fanboy and underdog going on, but if HP/Dell/etc decided to offer high end machines, AMD would actually get into serious trouble - because they'd run out of the high end chips, and be stuck with a bunch of midrange and low end chips they can do anything with, and unhappy customers all around. Enthusiasts would balk at paying $$$ for high end chips (think the old graphics card markups were bad during the cryptocurrency days?), and a ton of low end chips would flood the market no one wanted. (You could easily get 1050s and lower GPUs for regular price...)
Correct. OpenGL is an antique API - if you have a modern high end video card (like a GTX 1050 or higher) OpenGL will run like crap on it - it just has too much overhead causing most of the power to go underutilized.
The big problem is what OS X is going to use - Metal, Vulkan and the like all came out around the same time because of the issues of OpenGL One should note when Metal came out, Vulkan was actually AMD's API set - it was donated to Khronos to offer a standardized next-generation API set, and renamed to Vulkan. We are in a huge transition period where legacy apps will need to be ported over to take advantage of modern video card performance.
Let's see, he lost $180K, and FB went down 19% that one day, which would imply just under a million dollars or so in options trading.
Sorry, when you're dealing with those numbers ( and it's options, not real stock), it's somewhat hard to feel sorry anymore. Especially since if you read up on the guy and his celebrations of $300K in gains. $180K is a lot, but if you're doing that level of OPTIONS trading, you're probably fairly well taken care of and that $180k might just mean you skip today's Bentley purchase. (The people who do those sort of trades are either institutional, or already multimillionaire rich).
It's like people who complain a tax increase will cost them $50,000 more (happened). Turns out if you do that calculation, the guy was already doing about $2M in income so while the numbers are impressive, the real meaning behind them is hidden.
And yes, it always helps to do that calculation. Next time someone claims big numbers, find out what that tax increase actually means and you'll probably find out they're not going to be hurting quite so much.
First of all, cash isn't free to handle. If you're a small business, it might be close to free, but only because you don't have much of it. When you're handling a few thousands of dollars of cash daily, things get MUCH more interesting.
And it's easy to tell who's the "free" department - you can tell if a business simply turns off the register at the end of the day, or if they actually count the takes. If they don't count the takes, then in general they don't make enough cash (because it's not worth counting).
But once you do start raking in cash, you start to have to count it, which means you need to train employees how to handle cash. You need to tell them how to set up the register, how to ensure their takes are accurate, and how to sign off the register. This again is easy to see - when they run out of coins you note they always take a 410 roll of coins, but at the same time, they give the donator a $10 bill to equalize the transaction.(because every dollar has to be accounted for). And yes, when the register and cashbox do not agree, it's a very serious offense, minus some small percentage error (because people do realize short-changes and mistakes happen, but it shouldn't be more than a couple of bucks or so).
But you also have the problem - you have a few thousand dollars in cash now, and you have to deposit it at the bank. So someone has to go to the bank, wait in line, and make the deposit (which has to be counted, verified, etc). So that's someone's hour out of their day going to the bank, parking, lining up, etc., and hopefully not getting mugged or robbed along the way (another cost of handling cash). Or if you're lucky, you've got night depository permission so you can use those outside deposit boxes. But again, prime spot for being mugged (and it's after hours, too).
Hey, if you make enough money, perhaps you can call an armored car to take your cash for you to the bank, but that too is another cost for handling cash.
Oh, and let's not forget the whole robbery aspect - always a problem with cash-only businesses.
And finally, there's the problem of counterfeit bills. You might not think it, but large bills might be common - if you buy $50 worth of stuff, you would expect to get paid with $20 bills, $50 bills, and $100 bills. But $100 bills could be counterfeit, yet another cost (because counterfeit bills are non-redeemable, and it could've been changed with legitimate bills). Sure you can try to avoid taking in $50 or $100 bills, but if someone comes to you with $75 worth of purchase, they may not have $75 in smaller bills.
Now, the laws changed here recently (Canada) where retailers can charge the extra 3%, but it turns out most don't. Not because they can't, but because they realize that most of their transactions are credit and debit (which incur a fixed fee at least), and they fear losing that business. (And yes, businesses that were formerly cash-only have gone into accepting debt and later on, credit cards did see an uptick in business - both in number of customers through, and increases in size of transactions).
And to be honest, the ones that do charge, I tended to shop there less - about the ones that did were computer stores and they always did "cash discounted" pricing. It was always annoying to have to plan a visit to the bank to do a cash withdrawal in order to shop there, and they eventually lost me out to online shopping where the credit prices were included AND I had to pay shipping. (Not having to find parking in some obscure neighbourhood was a bonus, too).
Chances are, Charter was probably going to slip under the radar, then they started the union fight. The union reacted by filing a complaint with the PSC which led to an investigation and all the charges.
Charter was probably hoping to skirt by doing crap and having no one noticed, then they started fighting with the union and someone there realized Charter isn't quite on the up and up.
Usually it's just a pressure tactic, but I guess this time it backfired on Charter
That's only if MoviePass worked with theatres. Surprisingly, no theatre chain is working with them, and AMC blocked acceptance of MoviePass completely.
The thing is, MoviePass just has the wrong business model, period. Ticket prices, at least in the first two weeks, generally all go towards the studios. There is no way the theatre can offer MoviePass a break on ticket prices because of it.
The only time they can takes place on the third week or later, and that's when the theatre starts getting a small cut of the ticket revenue (generally to make up for smaller crowds and lower concession revenue).
It's generally fine - concessions bought during the first couple of weeks makes up for the loss in ticket revenue which is why theatres don't really mind giving 100% of ticket sales to the movie studios.
But even if there's an empty seat, as long as it's empty, the theatre pays $0. If that seat gets occupied, even if by someone with a discount ticket, the theatre must pay the full ticket price. So even a flop of a movie that doesn't sell in the second week, the theatre will rather show to an empty theatre (and pay $0) than have people come in.
This isn't an airline or transport business where an empty seat is a loss of potential revenue. An unsold ticket means an empty seat and it means the theatre does not owe the studio that money.
For MoviePass to survive, they must block the ability to view a movie in at least the first 14 days (the days the theatres make no money off tickets, and would rather have an empty seat and pay $0 to the studios for it over a discounted ticket where they'd have to make up the difference). They should also realize that concessions are how theatres make money, so offering free small popcorn or something to encourage sales of concessions would be necessary. (A small popcorn would lead to at least a drink sale at full price, and if people are going with friends, generally more concession sales as friends would want to buy stuff as well).
As for data mining, I'm sure the theatres have good ideas what their patrons do. "Dinner and a movie" is a common phrase for an activity, which is why they often have partners with restaurants. And after a movie is over, if it's a late showing, generally speaking most patrons are either going to go home, or to a bar. But in general, I don't think it's too valuable because too many people are doing too many different things.
And ensuring you have the latest and greatest hardware. I've never seen a chat client consume 30% CPU resources before, for what is effectively webchat. Other web chat things like Discord and such, no problem, they take less than 1% idle.
So if you're using Slack on anything lower than an i7 (i9 preferred) with 32GB of RAM, or preferentially a whole separate machine, it's painful.
And yes, webchat - because the "app" is actually a node.js/electron thing that wraps their webchat into an independent app. Which is preferable, since some browsers would just clog up if you used the web interface. The app just ran its own instance of Chromium independently so it wouldn't bog down your web browser.
Well, there's plenty of logging, and it's a bit cloudy too, so even if you joined today, you can still see previous chat logs (don't know how far back, but if you just keep scrolling back, more is fetched). So even if your boss bamboozles the new guy on the team, one he logs in to the chat, he can always scroll up and see what your conversation was.
Because of diminishing returns. It's called "Multithreading" and not "Multiprocessing" because even though it shows up as a separate CPU core, it's shared resources could benefit from threads of the same process far more than disparate threads, especially with respect to cache utilization.
It comes about because a modern CPU is super-scalar. That is, there are multiple execution units - often 3 (or more) integer units doing simple operations (adds, subtracts, bitshifts, boolean operations, etc), a couple of "complex" integer operations (multiply, divide), several load/store units (at least two each for loading and storing, and sometimes the effective address calculation uses the simple integer ALUs), multiple floating point units - again 2-3 for simple operations, 1 for complex operations and other units. Thing is, one program instruction stream, even with multiple instructions in flight simultaneously may not actually use all the resources. It's fairly cheap to tack on another front end execution unit and run another set of instructions through the same execution units, with the goal to have more utilization of the idle execution units.
Thus, disparate workloads generally work better - if one is more integer versus floating point, you can often run both together without them colliding. But if you're using all integer, and even worse, both are using the same units, then it will not speedup.
Good, because ZTE violated all sorts of regulations that were inposed on it.
The only reason Trump supported ZTE was China invested $500M into Trump hotels and granted Ivanka Trump a rare Chinese trademark. That's it. A little backroom dealing and let's forget that we asked you nicely to not resell those parts to places we asked you not to. And it's not like it was the first time, either.
If you ever wonder why Trump seems to care for Putin, again, remember Russia is one of the few places where Trump's business are actually succeeding, quite well in fact. Everywhere else they are doing terribly. All this from a President who has not divested himself of his businesses while in power.
Congress simply grew a backbone and enforced the order. Or shall we say, China did not scratch their backs obviously.
It depends. If the grame or application is programmed as a web game, it will likely be glitchy.
However, the new trend is WebAssembly, and you don't code in it directly, but use a cross-compiler. You write your code in C and it compiles down to a bunch of javascript (using a highly restricted subset that runs really fast) and your code runs. plugin-free.
The Internet Archive has plenty of play-in-browser DOSbox based games done like this. There are also in-browser versions of MUNT and other music players.
And the whole reason we have WebAssembly was someone at Mozilla wanted to have Unity be plugin-free (which created asm.js). So the future of web games is likely to be games that can be run natively or be run in a browser if a native runtime is not available.
Yes, this data is valuable because you can avoid the rush and get your pizza faster, hotter and fresher. During peak delivery periods, they would have to produce several orders at once for a single delivery run and thus where you lived determined how long the pizza is sitting in the delivery vehicle. Avoiding the peak periods means the kitchen makes the pizza and it's delivered straight to your door since you're the only order at the time.
I'm always curious - the tax scammers never identify who is in trouble. They always say "You're in big trouble with the tax man and the police are after you". Yet, most households in the US consist of multiple taxpayers - most families have both parents that work, and the kids may hold a job (summer or part time) as well.
Does it never occur to people that the scammers never identify the taxpayer involved? I admit once I nearly got fooled (it was a slick non-computerized speaker), but then I remembered the guy left a message without saying WHO was in trouble. If could be all of us, in which case they would've mentioned all of us, but they just said "you" which was highly ambiguous. Not like the tax man doesn't know how many taxpayers are at an address, and to actually identify which one they want to talk to.
Of course, I suppose I'm curious now what happens if they say that and I simply reply "I'm not eligible to pay taxes" and then ask who they want to speak to.
Finally, back on topic - you should mention the real legal process - the taxman always sends you a letter first, and legally, they cannot send the police after you, at least not without you completely knowing about it. As in there would be letters, there would be court service, court dates, there would be court sessions, judgements, etc. It will never be just a phone call. Justice is not swift, either - the process will take years. And even if at the end you're guilty, they generally still don't send the police after you.
Tax scammers use to only target the immigrant community (who were generally ignorant of how the legal system works, and in many countries outside North America, justice is often swift like that (which is what the scammers relied on). Here justice is slow and methodical and you get plenty of chances to defend yourself.
Easy, you only render the first frame and that's it. There were options to not play animated GIFs since browsers were first released, so all Firefox and other browsers need to do is simply modify the image renderer to only draw out the first frame. They can draw a simple play button to indicate that there are more images waiting to be drawn onto the page, but are currently blocked from doing so.
No one said you had to render out every frame of a multi-frame image. Just render the first, offer to play back other frames and leave it at that.
Amateur. YouTube has a few people who can get scammers on the phone for at least an hour, approaching 4 and a half. Though to be honest, a good chunk of it is simply dead air - they ask the victim to go out and buy gift cards or something (which is good for a couple of hours of scammers on hold).
And scammers have an "emergency out" - if they get exposed, they immediately lock the machine and syskey it to ensure it is unbootable. Of course, it all happens in a virtual machine, so restoring the damage is trivial, but it's funny.
Some YouTube channels:
Kitboga. He's had scammers insult him, make death threats and all sorts of fun.
This guy enacts revenge on the scammers by infecting their PC. Most of the scammers are just doing a script, so their PCs are often wide open and infectable, so it's possible to get RATs and such installed on their PC.
This guy investigates scammers computers. His latest video involves a bank login scam where he manages to install a RAT on the scammer's computer. He watches as a scammer attempts to register for and log into some elderly guy's bank account (luckily, the bank actually sends something to customers when this happens, so the scammer not only had no chance. He also alerted the bank who locked the online account for fraud.
Yes, it's fun watching scammer's computers and call centers get infected. Perhaps they can call themselves for tech support.
I admit, I get a lot of those calls. It's amazing how they always use the most robotic of voices. Also, apparently retailers are alert to the scam as well - several times they've been stopped when they see people buying thousands of dollars worth of iTunes or other gift cards. Though, someone was so embarrassed they made up a whole story about being "arrested" and transported and caused the police to issue an alert.
Except, the CPU was not overheating. The throttling did not happen because of the CPU
The CPU is powered by a Voltage Regulator Module (power delivery module). It turns out it is this chip that is overheating - when the CPU is going full tilt, it's demanding 125W from the VRM. This causes the VRMs to heat up and when they get close to their maximum thermal limits, they send a signal to the motherboard telling it to throttle the CPUs so they draw less power so the VRMs can cool down.
Part of thermal tuning is to adjust the CPU boost speeds such that it can boost to full speed, then throttle down slightly as the VRM and CPU heat up to a new max steady-state condition where the heat generated can be dissipated.
This is in part due to a documentation error in Intel's docs regarding max thermal power dissipation values.
The good news is if you tweak the throttle settings properly, you can keep the regulators from overheating, but the CPU still performing. This is what Apple did - they optimized the settings so the VRMs will not overheat and force a sudden throttling of the system. Doing this gives you a good 20% speed boost over the old models.
The bad news is if this was caught earlier so Apple could heatsink the VRMs to the CPU like they do with the GPU, you could get up to another 10-20% in performance because you can run the boosts longer since the VRMs would heat up slower.
The thing is, 99% of the CPU is idling even when running an IDE or Excel. So much so even a complex worksheet may take longer to open on a thin and light laptop, but in general, Excel or the IDE can crunch while I'm thinking about my next move.
It's important for the app to open fast and display what it knows. Recalculate in the background because by the time I navigate around to find what I want, the CPU is probably done with the calculating and macros and all that.
I love myself a beefy laptop, but while it's nice and powerful and runs everything I threw at it, it wasn't the best. Carrying it around was such a pain it was easier to just leave it at my desk. Traveling with it mean having to plan beforehand how I was going to carry it - its heft and weight made for real issues. And forget about using it enroute.
Today's modern processors mean it's Good Enough. It may not crank out the FPS as much, but eGPUs help, and it means when move about, I can carry my thin and light laptop with me - it slips in my bag and it's not that much heavier with the heft of other stuff. I can take it out and use it when I need to - granted, it's not got the graphical power so I can't play heavy games, but I can do plenty of other stuff with it. I can even code using an IDE with it. Getting decent battery life (something my laptop battlestation never had - the battery was more of a UPS than anything) is a bonus.
Granted, it's not for everybody - there are people who need to run 3D applications, Photoshop and other things that need raw power on the go and would need a huge heavy laptop. I just found while it was great, it was also more limiting. Turned out for me, I needed less a luggable, and more a portable.
Qualcomm makes parts. They sell parts. They do not sell modules. You cannot go to Qualcomm and buy a complete cellular module.
Instead, companies like Sierra Wireless, U-Blox, Widcomm and tons of others take Qualcomm's chips and turn them into modules of all sorts of form factors. Qualcomm provides support to those companies, while end users of those modules either use provided software by the module vendor, or write their own based on the 3GPP specification documents.
The demand is obvious, given the Nintendo NES Classic re-release (after that disasterous launch in 2006). Sure you can probably build a cheaper one, but again, Nintendo's version has all the ROMs licensed properly.
(And Nintendo didn't see Sega and Atari screwed themselves when they sold their stuff to AT Games, which is why those retro consoles never sold well - they may say 1000 games on the box, but it's really about 10 official real Sega Genesis or 2600 games, and 990 "inspired" games by AT Games themselves, so it's really a box full of shovelware). Whereas the NES Classic (and SNES Classic) contain nothing but real games.
And the re-release is done quietly this time around so it's actually possible to get them. Launch day stock lasted beyond the first day because unless you knew, it wasn't advertised.
Any why are we so quick to condemn actions like this, but cheer on the likes of people who go after GPL violators? it's exactly the same thing A GPL violator is a pirate - they've committed copyright infringement. Distributing GPL binaries without source is copyright violation - if you didn't agree to the GPL, your rights are what the law grants you.
Qualcomm is doing it because Apple does deliberately slow down the Qualcomm chipsets. Apple does it to minimize differences between phone models (there's not really "one" iPhone X, but a couple of them, and depending on the carrier, they get one or the other).
This is in part because of the Apple-Qualcomm lawsuit. Qualcomm obviously wants Apple to go all Qualcomm, but Apple doesn't want to be beholden to a single supplier - they've been down that road before (Motorola, IBM both failed to deliver the product Apple wanted on time, leading to many shortages in the past).
Samsung's doing it because their Galaxy S10 is coming out sooner or later, so they need to ramp up their advertising.
Of course, one wonders if the many Android bootlooping problems could be solved with Apple's "performance management" solution. I know it kicks in if the phone fails to boot because of the battery (i.e., what would've been a boot loop is then managed by the OS so instead of a dead rebooting phone, you have a working, but slower, phone). Some phones that bootloop were solved by simply reloading a kernel that disabled the high-power cores, a crude form of performance limiting, really.
It can be.
On Windows, where the default is Quick Removal, what happens is Windows disables write caching internally and doesn't let WriteFile() or Close() return until all the data buffers have been committed down to the media. This means when an application writes, the calls block (or the async I/O object is not signalled) until Windows is satisfied that it has pushed every byte of data down to the media. For USB devices, this means it sets the "Force Media Access" bit which tells the device to actually commit the data to the media. But, Windows assumes when the USB device replies, it's done (assuming no write caching done hidden from Windows). Windows is correct that as far as it can tell, it's done its job - the data has been pushed to hardware.
But as anyone can tell you, Windows 98 had a fatal bug where new PCs would shut down so quickly that despite Windows flushing all the caches, it would still shut down faster than the disks would finish flushing their caches to disk.
Similarly, if you know how SSDs work, you may want to wait a couple of seconds after the write completes so the controller can flush out the updated tables to media. Not doing so risks the controller tables being corrupted and drive nonfunctional, or the file being corrupted because of incorrect indices.
Finally, a bad Windows app can cause extensive wear on removable media because of this - if you write a small buffer, that buffer will go to media. If it writes another small buffer, then it forces another write
And on Linux and macOS, it's somewhat essential, since those OSes still enable write-back caching. Yanking the drive before unmounting will result in corrupt files or a corrupt filesystem.
Well, back before email, you typically sent a nice letter - handwritten or typed up and it was generally appreciated because it showed you took effort on it - after all, typing (or writing) a physical letter takes time, finding and envelope and paying for a stamp. It's why letters still generally have an effect.
But in this modern day of email, things are not so clear. Emails takes just a few minutes to type up, cost practically nothing to send, and may give an air of insincerity because of it. Especially when the recipient may be bombarded by hundreds of emails a day, so your note may get lost in the mess.
So sometimes a study like this helps clear the air that those little thank you notes that take almost no time at all to send are still appreciated. Of course, they still have to be personalized, so sending a mass emailed thank you note to everyone at your wedding thanking them for their gifts... may still come off as insincere.
Education is moving away from the 19th century module of facts and figures, and into the 21st century of analysis. It's moving very slowly, but it's become obvious it's the way to go.
Facts and figures were appropriate in the 19th century because it was hard to distribute, search and acquire information, so one could be more "educated" simply by knowing more. It's why tests are still a 19th century instrument of factual recall.
But today, we have facts and figures at our fingertips. There's no need to memorize facts and figures anymore, because it's all so easily looked up in an instant. Rote memorization isn't as useful.
Instead, the 21st century education system needs to go one further - given everyone has access to all the facts and figures, one needs to be able to analyze, evaluate and process them. The current education system is not set up to do this - to be able to take facts and figures, analyze them and figure out what's "fake news," "exaggerated," "contains omissions," or other errors. (Talking about real fabricated news here, not "news that puts Donald Trump in a less than positive light" "fake news"0.
That's unfortunately where college education generally comes in - in general, that's where the skills necessary to function in today's world comes in .