Not sure if that's because PHP is generally/inherently insecure, or just that most PHP developers don't know enough or care enough to make their apps secure.
Neither. It's mostly momentum. The problem is, critical software like MySQL was written without security in mind, and had poor support for query construction, leading to everyone rolling their own functionality, 100% of which contained security bugs, give or take some very small epsilon.
By the time they finally sorted out the security abomination that was the MySQL C API and replaced it with MySQLi, the old API was so pervasive that there are still sites running old, unsupported versions of PHP so that they can support the APIs, several years after they were removed in PHP 7. And I can pretty much guarantee that many of the "updated" versions just found ways to abuse the MySQLi API rather than doing things right.
So the newer, more modern API has been around for 13 years at this point, and vestiges of the old API still aren't 100% cleaned up, mostly because the PHP community started out neck-deep in crap code, and they're still shoveling their way out.
I think the reason this justifies an article is the comedy of errors that I've already highlighted. It's rather pathetic when the same company who maintains the search engine algorithms cannot properly police some of their own flagship product search results. One would think that a search like "office 2019 download" would be damn near hard-coded in Bing to provide the proper source, making it a hell of a lot harder for a pirate site to rise to the top of search results.
No, there should be a page that's popular enough and well-linked enough to get bumped to the highest position organically. If they have to hard-code hacks into their search engine just to get their products ranked high enough to beat the piracy sites, then they've already failed to properly market their product.
Even a non-hybrid SUV still gets 20+ MPH, which is about 15 cents a mile, on average, in the U.S. And even pure synthetic oil will cost less than $100 every 5,000 miles, which still adds less than two cents per mile. So even if you *massively* overestimate the total cost at 20 cents a mile, and even if your electrical power is completely free, it still takes 375,000 miles to cover the $75,000 difference between a Model X and a non-hybrid SUV.
Add in the extra several hundred dollars per year in vehicle licensing fees from the state of California, which gobble up approximately 100% of the savings in fuel costs, even when you assume that all your charging is free, and you're basically guaranteed to never break even, no matter how long you own the vehicle. (Okay, so we save a $40 smog check every two years after the first three, so we've got that going for us, which is nice, but....)
There are many reasons to buy a Model X. Saving money is not one of them, no matter how you look at it, unless you're in Europe where gasoline costs $7 per gallon.
For the Model 3, maybe. For an X, even with unlimited free supercharging, compared with a $27k RAV4 hybrid, the break-even point is still at a whopping 730,000 miles.
That's just it, though. Lithium is a commodity, and it is not that rare. Its price is set by the countries that mine the stuff, but the only thing preventing new mines from starting up is cost. Until the price hits certain points, some Lithium isn't worth bothering with. The moment it crosses certain thresholds, companies will start mining, and the price will stabilize. It won't go up forever.
BTW, the reason BEVs cost more is primarily economies of scale. Tesla's battery costs have come down fairly dramatically as their production volume has increased, in spite of rising Lithium prices. At this point, the packs cost somewhere around a third as much per kWh as they did eight years ago.
Be careful what you wish for: Tesla has cut every corner they could with the Model 3 (worker wellbeing, worker safety, fit and finish, repairability and repair affordability), and they still can't make the darn thing at the promised $36.000 price. Which is a high price for a car already. An all-EV future could be a future in you, average middle class person, cannot afford to drive.
Or a world in which the middle class demand to be paid better.
The average price of a new car, according to Experian, is $34,000. So a $36k car is only slightly above the average. Mind you, part of the reason is the crazy number of people buying $100k Teslas, but still....
Still has original battery, runs for 6 hours straight doing ANYTHING.
You obviously do not run Xcode. Or Finale. Or Lightroom. There's no way any of those apps runs anywhere close to 6 hours unless it is idle, and in the case of Finale, not even then.
The TouchBar isn't a colossal mistake. It is actually pretty damned useful, given the right Application. But I think I would have rather seen the TouchPad turned into a mini-Digitizer (with Pencil Support!), too.
The touch bar is a colossal mistake. Any apps in which you regularly have to use modifier keys with anything in the top row results in a crazy rate of false triggering that makes the app borderline unusable, and worse, option + touchbar actually opens the System Preferences app and kicks you entirely out of the app you're trying to use, so you have to immediately respond by hitting Command-Q just to get back to what you thought you were doing. In many cases, I've hit Command-Q dozens of times per minute because the false triggering on the touch bar is so bad.
And there's no way to actually disable the d**n touch bar in a single app, only globally. Even in normal use, I had so much false triggering that I disabled everything but volume, brightness, and escape on my personal machine, and all but brightness and escape on my work machine. I have a nearly blank strip. And even in that state, it is STILL a usability nightmare for me. Frankly, it makes me want to throw my MacBook Pro out a window on a near daily basis. I would gladly pay Apple an extra $2,000 right now if they would replace this piece of s**t top case with one containing a real keyboard, without otherwise changing the hardware. That's how much I hate the touch bar on a daily basis.
You don't get that kind of rage from a feature that isn't a colossal mistake. You just don't. If it had been an option, that would be okay. Unfortunately, choice isn't the Apple way. They get it in their heads that something is revolutionary and amazing, and they can't even begin to conceive of the possibility that others may use their computers differently than they do, and that their designs might result in an epic train wreck like the touch bar.
And lest you think I'm alone in this, I'm not. Nearly every single one of my immediate coworkers who ended up with a touch bar in their hardware refresh has hated it. One didn't mind it, but even he admitted that he got a lot of false triggering. Nobody thinks it works well. Zero. Not one single person out of at least a dozen software engineers.
So when I say that it was a colossal mistake, I say that based on a mountain of real-world user anecdotes from a lot of users. That's not to say that they couldn't redesign it in a way that it will work well — add extra separation between it and the keyboard, insert a row of function keys in the middle, include sensors in the top row of keys to turn off the touchbar, etc. — but as implemented, it is, IMO, a complete disaster like nothing I've ever seen from Apple. This makes the Pippin look like an amazing gaming experience. This makes the Cube look like a first-class desktop. This makes the hockey puck mouse seem ergonomically correct. Need I continue?
I think that it would be silly to say that Apple didn't run to the mobile device market pretty quickly, and pretty whole-hog; but I think that with the move to the new HQ, they have reapportioned their talent so that they can start to chew gum and walk at the same time again. The new Mac mini was pretty close to perfect (assuming you are ok with adding an eGPU for graphic-intensive stuff), and that you can effortlessly put together a little (literally!) rendering/build-cluster of them for CHEEEEP, and, IMHO, shows real promise for new Mac hardware to come in this coming year.
The new Mini is okay. The problem is that for about four years, the Mini was a useless joke, and the replacement still doesn't have as much internal storage as you could build out in the 2012 model at the time, much less as much as you could stick in the 2012 model today. Mine has 2 TB of internal storage, and I expected a new model to at least be an upgrade over that, capacity-wise, but it isn't.
1. Provide an IP. Might eventually need to start doing v6 rather than v4, but that is a municipal task.
No, that would all be an ISP task. The municipality leases out a physical glass fiber that goes from a government building (the central office) to the customer site. All equipment involved in using that fiber, up to and including provisioning IPs, routes, converting from optical to Ethernet or whatever, etc. is the responsibility of the ISP leasing the glass fiber from the municipality.
2. Services like local channels and even cable are presumably delivered via multicast, with perhaps some encryption for security.
Again, entirely determined by the cable company providing service. More than likely, they would use CATV-over-fiber, but they could also use anything from multicast to a head end that provides unicast streams to individual Roku boxes owned or leased by the customer. And because a cable company could put their antenna tower and satellite dishes anywhere within the covered region and potentially send the entire chunk of data digitally to their head end in the CO via a single leased fiber, the cost of setting up a new cable company would be roughly the cost of tower and satellite dish installation plus head end equipment and receivers, and that's it. So municipal fiber opens up the possibility of real competition in cable companies, too, which is mostly impractical in a world where everybody has to run their own coax.
3. Basically all the ISPs that show up to the CO are basically valid routes to the rest of the internet. The routers at the ISP deal with those details, providing you access to 1 or more valid routes. This could potentially be handled even for large buildings since your connecting an IP to being able to access certain routers and routes.
Presumably, the owner of the building would get service and distribute it to everyone in the building, but other options certainly would also be possible. That's mostly between your landlord and whatever ISP the landlord decides to work with.:-)
4. Phone is pretty much just whatever IP phone service you want, save those packets might have a higher priority.
5. Netflix could install its own box or boxes right in the CO to eliminate the bullshit, so it becomes its own ISP, sort of.
In all likelihood, they would provide a single caching box that would be tied to all of the various ISPs at that location, obtaining its upstream service for pulling down new content from whichever ISP offered a better deal.
Check any grid's total load and you will see its peak daily at 8pm.
Actually, no. Here's California's grid for the middle of the summer: 07/11/2018. The peak is at 5 p.m., which is because you're starting to ramp up home power use while ramping down business power use. Demand peaks later on cooler days, but it also peaks lower, because there's less power being used to cool.
Either way, though, what you see is that power starts ramping up significantly as the day gets hotter, lagging behind temperature at both ends. It starts to exceed the average daily consumption at eleven or twelve in the morning and goes up from there. Is there a perfect match between solar power and cooling needs? No. Are they reasonably correlated? Yes.
Also, we frack for natural gas.
No, we frack for oil. The overwhelming majority of natural gas comes as a side effect of oil exploration. I'm sure that a very small percentage of wells produce only natural gas, just as a very small percentage of oil wells produce essentially zero natural gas. However, these are by far the exception, not the rule. Nearly all natural gas wells also produce oil and vice versa.
Yes, batteries have gotten a bit better since the Pismo days (was it even LiOn?). Nope, it was NiCD (!!!) (I think it was actually at least NiMH in 2000):
Apple switched from NiCd to NiMh in the second generation of PowerBook 5300 batteries (circa 1995), and from NiMh to Lithium Ion (round cell) in the PowerBook 3400 (1997). Note that they were round-cell Lithium ion, not Lithium polymer (flat bags). Those didn't come until much later.
But I'm not sure even making the MBP as thick as a Mid-2012 Non-Retina version (which I happen to own), would make room for 2 100 WH batteries
The retina display is considerably thinner than the pre-retina display was, so a retina MBP that's as thick as the pre-retina MBP would have a thicker body than the pre-retina MBP. I'm not sure you could actually squeeze in two 100 Wh batteries, but you could obviously do a lot better than what we have right now.
Sorry, making MBPs out of a solid chunk of Aluminum makes for a very rugged laptop; but unfortunately, has kinda killed the idea of removable "drive bays", regardless of the thickness.
Solid block? The bottom is a separate plate that screws on. If done right, removable bottom batteries shouldn't significantly weaken the hardware structurally, I wouldn't think. Side batteries would be harder, but oh, so cool if they could pull it off.
I guess the best thing to do is invest in one of those externals, like this Poster [slashdot.org] suggests. Is it a perfect solution?
I have such a device. In practice, I rarely use it, because there's not a huge difference between being tethered to a large external brick and being tethered to a wall outlet, in my experience. It's still pretty much the opposite of portable.
Is anything? And just remember how much faster, and how much more I/O expandable, that 2016-2018 MBP is than your beloved Pismo...;-)
Yeah. Now I just need a separate laptop bag for all the dongles required to actually use that expandability.:-D
You do realize, of course, that Apple has been running more Mac ads than iPhone ads ever since the beginning of November. I have seen two ads: The MacBook Air ad, and now followed by the black and white ad showing people using Macs in various settings (including a shot of Paul McCartney behind a mixing desk).
Why would they advertise products that are selling well? There's no need to advertise the iPhone. But the Mac platform is a train wreck, largely from having been ignored so thoroughly for so long, and with most of the "improvements" being colossal mistakes like the touch bar, so much so that sometimes I think the best thing Apple could do would be to open source macOS under a permissive license and cede the computer market to companies that actually still care about building computers.
IMO, it is pretty obvious when Apple stopped caring about Macs. Things started slipping in about 2008 or 2009, and the whole platform has been going downhill ever since. The last truly good version of OS X was 10.6.8, released in '09. The last XServe was released in 2009. The last non-speed-bump cheese grater Mac Pro was released in 2009.
To be fair, the annual release cadence (driven largely by trying to keep up with iOS, which ties major OS versions to iPhone hardware releases) is a big part of why macOS sucks so badly now, but Lion wasn't exactly a polished release, either. Basically, iPhone was simultaneously the best and worst thing that has happened to Apple since Steve's second coming. It was great for the stock. It was terrible for the Mac platform.
What's really funny is that the same thing happened to Apple back before S.J. returned, only it was the Newton that sucked all the resources away from the Mac. One of the smartest things he did was to kill that product line and save the Mac. These days, Apple lacks focus, and I'm not entirely convinced that they are capable of being both a computer company and a toy phone/tablet company. Maybe they should pick one, open source the other or license it to one or more of their competitors, and hope for the best.
I mean, who wouldn't want macOS to be a competitor to Windows on everything from Dell to HP to Asus?
That's because everyone but Apple lies like a dog even worse than Apple does when it comes to battery life claims.
FTFY. I don't think I've ever gotten anywhere close to ten hours on my retina MBP. On average, I'm lucky to get much more than three hours unless I'm doing something that uses almost zero CPU, like web browsing.
The problem is, power management is not a replacement for a larger battery, but unfortunately, Apple's hardware engineering managers, with their utterly myopic focus on making laptops thinner, can't seem to let that reality seep through their thick skulls. So instead of giving us the maximum battery size you can legally carry on an airplane (100 Wh), each generation of MBP has had a smaller battery than the one before it. Therefore, in my experience, actual battery life has gotten measurably worse every time I've upgraded my hardware.
Of all the Macs I've owned, the one with the best battery life was the PowerBook Pismo, way back at the turn of the century. Why? Because the pre-iPhone Apple understood that having removable batteries means you can have more than one, and that what matters is not the best-case battery life, which most users will never actually see, but rather the worst-case battery life, which all users will at least sometimes see. The 4x difference between best-case and worst-case is a real kick in the teeth, and will continue to be until such time as the worst-case battery life improves by at least a factor of two.
The Pismo, in particular, was notable in that it had two battery bays, each of which could hold a roughly 70 Wh battery. If needed, I could easily carry around a third battery and hot-swap it for the fully-drained battery without even putting the laptop to sleep. The result was a whopping 9+ hours of real-world battery life (best-case 15 hours) even while running apps like Photoshop or audio editing software. Every laptop Apple has made since then has been a complete joke by comparison, unless you're using the laptop for a task that an iPad can handle just as well, such as light-duty web browsing.
These days, I always carry a power supply around, and assume that if I'm doing anything even remotely interesting for more than an hour or so, I'm going to end up tethered to a wall outlet. Gone are the days of writing software on the beach. Modern Apple hardware just can't do it anymore. Neither can anybody else's, to be fair, but Apple is pretty just about the only company whose hardware ever could, and I miss that.
Mind you, I don't relish going back to the thickness of the Pismo (mainly thick because of the design of the plastic case and the use of round cells in the battery pack), but I would gladly go back to at least the thickness of the pre-retina MBP if it got me two removable 99.9 Wh LiPo batteries.
Given the battery and CPU specs, I would guess more like 6–7 hours of light web browsing, 1 hour to 75 minutes under heavy CPU/GPU load.
That said, I agree that the battery looks massively under-specified for a 17" laptop, unless I'm missing something (such as deliberately massively throttled CPU performance).
I suspect the marketing is being laid on pretty thick here. They're claiming that a 72 Wh battery will run this 17" laptop for up to 19.5 hours, while the 15" MacBook Pro's 83.6 Wh battery is rated for up to 10 hours. I know it's a different CPU, but I have a hard time believing that Intel made a 2x improvement in performance-per-watt without a die shrink, much less enough extra improvement beyond that to make up for the extra power required by a significantly larger screen.
I think you may be missing his point. I've had SSD's die on me as well with absolutely no warning. What's unnerving about it is you have no idea why it failed. Good engineers like failure analysis; it helps determine if you're buying a crappy product, running your product out of spec, or any number of other metrics which can inform future purchases.
Statistically, without even knowing what the particular product was, I can tell you what caused it: RoHS.
The change from lead-based solder to lead-free solder is one of the major causes of premature electronics failures — probably more common than all other causes put together. Between tin whiskers, cold solder joints, and stress fractures caused by thermal expansion of component packages, the RoHS lead-free solder rule is a clear example of environmentalism gone amok. Instead of improving our environment by reducing the amount of lead going out into the world, it has, IMO, made our environment worse by dramatically increasing the amount of hardware discarded as junk long before it otherwise would have been.
It's usually because of the controller or RAM / Cache errors in processing that corrupts the firmware or dynamic LBA flash block allocation table (database). This renders the reset of the NAND flash partially or totally inaccessible. Quality "prosumer" drives are supposed to have extra hardware (capacitance) to prevent half-writes upon a dirty shutdown (abrupt loss of power). But regardless, any corruption on write-back can render the drive "bricked".
And by this, you mean that some really bad SSD manufacturers still haven't learned the concept of log-structured storage. The problem of handling a partial write was solved a couple of decades ago. You roll back the partial transaction to the last checkpoint, then say, "whoops, that write never happened".
Basically, in addition to a flat mapping table (as a cache), you store a copy of the mapping table (a checkpoint) with modifications in a log format. Each time you power on the drive, it ignores the cached flat mapping table (if it even bothers to persist it to disk), and reads the last checkpoint table, then replays the transaction log after that. When it reaches the last completed transaction in the log, it now has a valid mapping table that it is up-to-date to the maximum extent possible. A write operation is considered committed as soon as the transaction is added to the log, and existing used space is not reclaimed until that log write has occurred, ensuring that every write is effectively an atomic operation. Periodically, you write out a new flat table as a checkpoint, and after ensuring that it has been fully written, you then mark the oldest checkpoint and associated log pages as free for reuse.
We were talking about this back when I was in grad school, around the turn of the century, precisely to prevent those sorts of failures. So IMO, if any SSD manufacturer still isn't doing a transactional/log-based mapping table between blocks and flash pages at this point, their hardware isn't good enough to use for storing system logs for a flush toilet, much less critical data. I mean, this is really *basic* stuff, and has been the norm for at least a decade.
For the same reasons you don't want the town hall to run pizzeria.
Funny, nearly every place that has tried it has had good results. The key, of course, is to do it correctly. First, the city comes in and installs the lines. Then, they contract out service to local ISPs. ISPs compete with each other to provide service over the existing infrastructure, which means that the cost for a new ISP to join the mix is negligible, leading to a highly competitive market even in low-population areas. And the only thing the municipality has to do is maintain the infrastructure, which, it turns out, is something that government does pretty well.
That's somewhat disingenuous. Most natural gas comes along with oil that is being pulled out of the ground anyway. Thus, even when you factor in some leakage, the environmental impact of natural gas is still remarkably close to zero, because the only real alternative to using it as a power source is burning it off at the wellhead or releasing it into the atmosphere.
And the claim that solar and wind inherently require more natural gas is also pretty much absurd. Solar power is at its strongest when power usage is at its highest — during the hottest parts of the hottest days. Solar power is a great choice to balance out the highly time-skewed loads caused by air conditioning during much of the year. And even after you pass that balancing point, you're still reducing the environmental impact of non-renewable fuels.
You're are, of course, cutting into the ability to use nuclear as base load at night; everything else can ramp pretty quickly as needed to balance out the wind and solar variation, but nuclear can't. But that's mostly an orthogonal discussion. And it remains to be seen whether newer nuclear plant designs can eliminate those limitations.
1. The existing grid cannot support a full conversion to electric cars, in fact estimates are that it will start to fail at less than 25%.
The grid almost always isn't the problem. Inadequate power generation is.
2. Home solar for charging? Pull the other one. Unless you drive your electric vehicle very short distances occasionally there is no easy you will have that kind of capacity
The average home solar setup is 5 kW. A full charge on a Tesla (~300 miles) would require about 20 full-sun-equivalent hours on such a system, which in most of the U.S., on average, takes about four days. This leaves almost half of the produced power to balance out other home electricity use. Mind you, an EV owner is likely to want a larger-than-average system, but we aren't talking about orders of magnitude here. A factor of two will provide enough power for two EVs driving 300 miles per week, which is above-average mileage.
(and that's ignoring the fact that most EVs are charging at night, when solar is... Less that efficient shall we say?)
During the day, you're putting power back into the grid and getting paid for production, because otherwise companies would have to burn fossil fuels to produce that power. At night, you're buying surplus power that is no longer needed for cooling homes and businesses, and using it to charge your cars. The exact environmental impact, of course, depends on your base load mix.
Oh, I thought for a minute that they were partnering with a Motown singing group. It kind of sounds like Samsung tried to hurry love. You can't do that, you know.
Slashdot Mirror
Neither. It's mostly momentum. The problem is, critical software like MySQL was written without security in mind, and had poor support for query construction, leading to everyone rolling their own functionality, 100% of which contained security bugs, give or take some very small epsilon.
By the time they finally sorted out the security abomination that was the MySQL C API and replaced it with MySQLi, the old API was so pervasive that there are still sites running old, unsupported versions of PHP so that they can support the APIs, several years after they were removed in PHP 7. And I can pretty much guarantee that many of the "updated" versions just found ways to abuse the MySQLi API rather than doing things right.
So the newer, more modern API has been around for 13 years at this point, and vestiges of the old API still aren't 100% cleaned up, mostly because the PHP community started out neck-deep in crap code, and they're still shoveling their way out.
No, there should be a page that's popular enough and well-linked enough to get bumped to the highest position organically. If they have to hard-code hacks into their search engine just to get their products ranked high enough to beat the piracy sites, then they've already failed to properly market their product.
FTFY.
And it'll be great, really great, I tell you — so great you won't believe how unbelievably great this wall will be.
At $35k, maybe. Not at $100k.
Even a non-hybrid SUV still gets 20+ MPH, which is about 15 cents a mile, on average, in the U.S. And even pure synthetic oil will cost less than $100 every 5,000 miles, which still adds less than two cents per mile. So even if you *massively* overestimate the total cost at 20 cents a mile, and even if your electrical power is completely free, it still takes 375,000 miles to cover the $75,000 difference between a Model X and a non-hybrid SUV.
Add in the extra several hundred dollars per year in vehicle licensing fees from the state of California, which gobble up approximately 100% of the savings in fuel costs, even when you assume that all your charging is free, and you're basically guaranteed to never break even, no matter how long you own the vehicle. (Okay, so we save a $40 smog check every two years after the first three, so we've got that going for us, which is nice, but....)
There are many reasons to buy a Model X. Saving money is not one of them, no matter how you look at it, unless you're in Europe where gasoline costs $7 per gallon.
For the Model 3, maybe. For an X, even with unlimited free supercharging, compared with a $27k RAV4 hybrid, the break-even point is still at a whopping 730,000 miles.
That's just it, though. Lithium is a commodity, and it is not that rare. Its price is set by the countries that mine the stuff, but the only thing preventing new mines from starting up is cost. Until the price hits certain points, some Lithium isn't worth bothering with. The moment it crosses certain thresholds, companies will start mining, and the price will stabilize. It won't go up forever.
BTW, the reason BEVs cost more is primarily economies of scale. Tesla's battery costs have come down fairly dramatically as their production volume has increased, in spite of rising Lithium prices. At this point, the packs cost somewhere around a third as much per kWh as they did eight years ago.
Or a world in which the middle class demand to be paid better.
The average price of a new car, according to Experian, is $34,000. So a $36k car is only slightly above the average. Mind you, part of the reason is the crazy number of people buying $100k Teslas, but still....
You obviously do not run Xcode. Or Finale. Or Lightroom. There's no way any of those apps runs anywhere close to 6 hours unless it is idle, and in the case of Finale, not even then.
The touch bar is a colossal mistake. Any apps in which you regularly have to use modifier keys with anything in the top row results in a crazy rate of false triggering that makes the app borderline unusable, and worse, option + touchbar actually opens the System Preferences app and kicks you entirely out of the app you're trying to use, so you have to immediately respond by hitting Command-Q just to get back to what you thought you were doing. In many cases, I've hit Command-Q dozens of times per minute because the false triggering on the touch bar is so bad.
And there's no way to actually disable the d**n touch bar in a single app, only globally. Even in normal use, I had so much false triggering that I disabled everything but volume, brightness, and escape on my personal machine, and all but brightness and escape on my work machine. I have a nearly blank strip. And even in that state, it is STILL a usability nightmare for me. Frankly, it makes me want to throw my MacBook Pro out a window on a near daily basis. I would gladly pay Apple an extra $2,000 right now if they would replace this piece of s**t top case with one containing a real keyboard, without otherwise changing the hardware. That's how much I hate the touch bar on a daily basis.
You don't get that kind of rage from a feature that isn't a colossal mistake. You just don't. If it had been an option, that would be okay. Unfortunately, choice isn't the Apple way. They get it in their heads that something is revolutionary and amazing, and they can't even begin to conceive of the possibility that others may use their computers differently than they do, and that their designs might result in an epic train wreck like the touch bar.
And lest you think I'm alone in this, I'm not. Nearly every single one of my immediate coworkers who ended up with a touch bar in their hardware refresh has hated it. One didn't mind it, but even he admitted that he got a lot of false triggering. Nobody thinks it works well. Zero. Not one single person out of at least a dozen software engineers.
So when I say that it was a colossal mistake, I say that based on a mountain of real-world user anecdotes from a lot of users. That's not to say that they couldn't redesign it in a way that it will work well — add extra separation between it and the keyboard, insert a row of function keys in the middle, include sensors in the top row of keys to turn off the touchbar, etc. — but as implemented, it is, IMO, a complete disaster like nothing I've ever seen from Apple. This makes the Pippin look like an amazing gaming experience. This makes the Cube look like a first-class desktop. This makes the hockey puck mouse seem ergonomically correct. Need I continue?
The new Mini is okay. The problem is that for about four years, the Mini was a useless joke, and the replacement still doesn't have as much internal storage as you could build out in the 2012 model at the time, much less as much as you could stick in the 2012 model today. Mine has 2 TB of internal storage, and I expected a new model to at least be an upgrade over that, capacity-wise, but it isn't.
1. Provide an IP. Might eventually need to start doing v6 rather than v4, but that is a municipal task.
No, that would all be an ISP task. The municipality leases out a physical glass fiber that goes from a government building (the central office) to the customer site. All equipment involved in using that fiber, up to and including provisioning IPs, routes, converting from optical to Ethernet or whatever, etc. is the responsibility of the ISP leasing the glass fiber from the municipality.
2. Services like local channels and even cable are presumably delivered via multicast, with perhaps some encryption for security.
Again, entirely determined by the cable company providing service. More than likely, they would use CATV-over-fiber, but they could also use anything from multicast to a head end that provides unicast streams to individual Roku boxes owned or leased by the customer. And because a cable company could put their antenna tower and satellite dishes anywhere within the covered region and potentially send the entire chunk of data digitally to their head end in the CO via a single leased fiber, the cost of setting up a new cable company would be roughly the cost of tower and satellite dish installation plus head end equipment and receivers, and that's it. So municipal fiber opens up the possibility of real competition in cable companies, too, which is mostly impractical in a world where everybody has to run their own coax.
3. Basically all the ISPs that show up to the CO are basically valid routes to the rest of the internet. The routers at the ISP deal with those details, providing you access to 1 or more valid routes. This could potentially be handled even for large buildings since your connecting an IP to being able to access certain routers and routes.
Presumably, the owner of the building would get service and distribute it to everyone in the building, but other options certainly would also be possible. That's mostly between your landlord and whatever ISP the landlord decides to work with. :-)
4. Phone is pretty much just whatever IP phone service you want, save those packets might have a higher priority.
5. Netflix could install its own box or boxes right in the CO to eliminate the bullshit, so it becomes its own ISP, sort of.
In all likelihood, they would provide a single caching box that would be tied to all of the various ISPs at that location, obtaining its upstream service for pulling down new content from whichever ISP offered a better deal.
Actually, no. Here's California's grid for the middle of the summer: 07/11/2018. The peak is at 5 p.m., which is because you're starting to ramp up home power use while ramping down business power use. Demand peaks later on cooler days, but it also peaks lower, because there's less power being used to cool.
Either way, though, what you see is that power starts ramping up significantly as the day gets hotter, lagging behind temperature at both ends. It starts to exceed the average daily consumption at eleven or twelve in the morning and goes up from there. Is there a perfect match between solar power and cooling needs? No. Are they reasonably correlated? Yes.
No, we frack for oil. The overwhelming majority of natural gas comes as a side effect of oil exploration. I'm sure that a very small percentage of wells produce only natural gas, just as a very small percentage of oil wells produce essentially zero natural gas. However, these are by far the exception, not the rule. Nearly all natural gas wells also produce oil and vice versa.
Apple switched from NiCd to NiMh in the second generation of PowerBook 5300 batteries (circa 1995), and from NiMh to Lithium Ion (round cell) in the PowerBook 3400 (1997). Note that they were round-cell Lithium ion, not Lithium polymer (flat bags). Those didn't come until much later.
The retina display is considerably thinner than the pre-retina display was, so a retina MBP that's as thick as the pre-retina MBP would have a thicker body than the pre-retina MBP. I'm not sure you could actually squeeze in two 100 Wh batteries, but you could obviously do a lot better than what we have right now.
Solid block? The bottom is a separate plate that screws on. If done right, removable bottom batteries shouldn't significantly weaken the hardware structurally, I wouldn't think. Side batteries would be harder, but oh, so cool if they could pull it off.
I have such a device. In practice, I rarely use it, because there's not a huge difference between being tethered to a large external brick and being tethered to a wall outlet, in my experience. It's still pretty much the opposite of portable.
Yeah. Now I just need a separate laptop bag for all the dongles required to actually use that expandability. :-D
Ah. Those were the machine dimensions. I misread that being as the dimensions of the display. Then yes, those numbers make complete sense.
Why would they advertise products that are selling well? There's no need to advertise the iPhone. But the Mac platform is a train wreck, largely from having been ignored so thoroughly for so long, and with most of the "improvements" being colossal mistakes like the touch bar, so much so that sometimes I think the best thing Apple could do would be to open source macOS under a permissive license and cede the computer market to companies that actually still care about building computers.
IMO, it is pretty obvious when Apple stopped caring about Macs. Things started slipping in about 2008 or 2009, and the whole platform has been going downhill ever since. The last truly good version of OS X was 10.6.8, released in '09. The last XServe was released in 2009. The last non-speed-bump cheese grater Mac Pro was released in 2009.
To be fair, the annual release cadence (driven largely by trying to keep up with iOS, which ties major OS versions to iPhone hardware releases) is a big part of why macOS sucks so badly now, but Lion wasn't exactly a polished release, either. Basically, iPhone was simultaneously the best and worst thing that has happened to Apple since Steve's second coming. It was great for the stock. It was terrible for the Mac platform.
What's really funny is that the same thing happened to Apple back before S.J. returned, only it was the Newton that sucked all the resources away from the Mac. One of the smartest things he did was to kill that product line and save the Mac. These days, Apple lacks focus, and I'm not entirely convinced that they are capable of being both a computer company and a toy phone/tablet company. Maybe they should pick one, open source the other or license it to one or more of their competitors, and hope for the best.
I mean, who wouldn't want macOS to be a competitor to Windows on everything from Dell to HP to Asus?
Displays are measured corner to corner. Of course, sqrt(15^2 + 10.5^2) is actually 18.3 inches, so some of those numbers must be off.
If it's a 16:10 display, to compute the dimensions, use w = 1.6h and fill that into the pythagorean theorem (c^2 = a^2 + b^2).
(17^2) = [(1.6h)^2 + h^2] = [2.56(h^2) + h^2] = 3.56(h^2)
Thus, h^2 = 17^2/3.56, and h ~= 9", w ~= 14.4". Yeah, those numbers are way, way off. Either that or it isn't a 17" screen.
FTFY. I don't think I've ever gotten anywhere close to ten hours on my retina MBP. On average, I'm lucky to get much more than three hours unless I'm doing something that uses almost zero CPU, like web browsing.
The problem is, power management is not a replacement for a larger battery, but unfortunately, Apple's hardware engineering managers, with their utterly myopic focus on making laptops thinner, can't seem to let that reality seep through their thick skulls. So instead of giving us the maximum battery size you can legally carry on an airplane (100 Wh), each generation of MBP has had a smaller battery than the one before it. Therefore, in my experience, actual battery life has gotten measurably worse every time I've upgraded my hardware.
Of all the Macs I've owned, the one with the best battery life was the PowerBook Pismo, way back at the turn of the century. Why? Because the pre-iPhone Apple understood that having removable batteries means you can have more than one, and that what matters is not the best-case battery life, which most users will never actually see, but rather the worst-case battery life, which all users will at least sometimes see. The 4x difference between best-case and worst-case is a real kick in the teeth, and will continue to be until such time as the worst-case battery life improves by at least a factor of two.
The Pismo, in particular, was notable in that it had two battery bays, each of which could hold a roughly 70 Wh battery. If needed, I could easily carry around a third battery and hot-swap it for the fully-drained battery without even putting the laptop to sleep. The result was a whopping 9+ hours of real-world battery life (best-case 15 hours) even while running apps like Photoshop or audio editing software. Every laptop Apple has made since then has been a complete joke by comparison, unless you're using the laptop for a task that an iPad can handle just as well, such as light-duty web browsing.
These days, I always carry a power supply around, and assume that if I'm doing anything even remotely interesting for more than an hour or so, I'm going to end up tethered to a wall outlet. Gone are the days of writing software on the beach. Modern Apple hardware just can't do it anymore. Neither can anybody else's, to be fair, but Apple is pretty just about the only company whose hardware ever could, and I miss that.
Mind you, I don't relish going back to the thickness of the Pismo (mainly thick because of the design of the plastic case and the use of round cells in the battery pack), but I would gladly go back to at least the thickness of the pre-retina MBP if it got me two removable 99.9 Wh LiPo batteries.
Given the battery and CPU specs, I would guess more like 6–7 hours of light web browsing, 1 hour to 75 minutes under heavy CPU/GPU load.
That said, I agree that the battery looks massively under-specified for a 17" laptop, unless I'm missing something (such as deliberately massively throttled CPU performance).
I suspect the marketing is being laid on pretty thick here. They're claiming that a 72 Wh battery will run this 17" laptop for up to 19.5 hours, while the 15" MacBook Pro's 83.6 Wh battery is rated for up to 10 hours. I know it's a different CPU, but I have a hard time believing that Intel made a 2x improvement in performance-per-watt without a die shrink, much less enough extra improvement beyond that to make up for the extra power required by a significantly larger screen.
Statistically, without even knowing what the particular product was, I can tell you what caused it: RoHS.
The change from lead-based solder to lead-free solder is one of the major causes of premature electronics failures — probably more common than all other causes put together. Between tin whiskers, cold solder joints, and stress fractures caused by thermal expansion of component packages, the RoHS lead-free solder rule is a clear example of environmentalism gone amok. Instead of improving our environment by reducing the amount of lead going out into the world, it has, IMO, made our environment worse by dramatically increasing the amount of hardware discarded as junk long before it otherwise would have been.
And by this, you mean that some really bad SSD manufacturers still haven't learned the concept of log-structured storage. The problem of handling a partial write was solved a couple of decades ago. You roll back the partial transaction to the last checkpoint, then say, "whoops, that write never happened".
Basically, in addition to a flat mapping table (as a cache), you store a copy of the mapping table (a checkpoint) with modifications in a log format. Each time you power on the drive, it ignores the cached flat mapping table (if it even bothers to persist it to disk), and reads the last checkpoint table, then replays the transaction log after that. When it reaches the last completed transaction in the log, it now has a valid mapping table that it is up-to-date to the maximum extent possible. A write operation is considered committed as soon as the transaction is added to the log, and existing used space is not reclaimed until that log write has occurred, ensuring that every write is effectively an atomic operation. Periodically, you write out a new flat table as a checkpoint, and after ensuring that it has been fully written, you then mark the oldest checkpoint and associated log pages as free for reuse.
We were talking about this back when I was in grad school, around the turn of the century, precisely to prevent those sorts of failures. So IMO, if any SSD manufacturer still isn't doing a transactional/log-based mapping table between blocks and flash pages at this point, their hardware isn't good enough to use for storing system logs for a flush toilet, much less critical data. I mean, this is really *basic* stuff, and has been the norm for at least a decade.
Funny, nearly every place that has tried it has had good results. The key, of course, is to do it correctly. First, the city comes in and installs the lines. Then, they contract out service to local ISPs. ISPs compete with each other to provide service over the existing infrastructure, which means that the cost for a new ISP to join the mix is negligible, leading to a highly competitive market even in low-population areas. And the only thing the municipality has to do is maintain the infrastructure, which, it turns out, is something that government does pretty well.
That's somewhat disingenuous. Most natural gas comes along with oil that is being pulled out of the ground anyway. Thus, even when you factor in some leakage, the environmental impact of natural gas is still remarkably close to zero, because the only real alternative to using it as a power source is burning it off at the wellhead or releasing it into the atmosphere.
And the claim that solar and wind inherently require more natural gas is also pretty much absurd. Solar power is at its strongest when power usage is at its highest — during the hottest parts of the hottest days. Solar power is a great choice to balance out the highly time-skewed loads caused by air conditioning during much of the year. And even after you pass that balancing point, you're still reducing the environmental impact of non-renewable fuels.
You're are, of course, cutting into the ability to use nuclear as base load at night; everything else can ramp pretty quickly as needed to balance out the wind and solar variation, but nuclear can't. But that's mostly an orthogonal discussion. And it remains to be seen whether newer nuclear plant designs can eliminate those limitations.
The grid almost always isn't the problem. Inadequate power generation is.
The average home solar setup is 5 kW. A full charge on a Tesla (~300 miles) would require about 20 full-sun-equivalent hours on such a system, which in most of the U.S., on average, takes about four days. This leaves almost half of the produced power to balance out other home electricity use. Mind you, an EV owner is likely to want a larger-than-average system, but we aren't talking about orders of magnitude here. A factor of two will provide enough power for two EVs driving 300 miles per week, which is above-average mileage.
During the day, you're putting power back into the grid and getting paid for production, because otherwise companies would have to burn fossil fuels to produce that power. At night, you're buying surplus power that is no longer needed for cooling homes and businesses, and using it to charge your cars. The exact environmental impact, of course, depends on your base load mix.
Oh, I thought for a minute that they were partnering with a Motown singing group. It kind of sounds like Samsung tried to hurry love. You can't do that, you know.