Right, when you take a highly specialized niche and just put "X" in, then it sounds a lot better.
You seem to be denying that there would be any sort of subject where "asking a question" in the literal sense would be a useful way to gain knowledge. But there are in fact subjects that have too much required knowledge for that to work; you have to actually study the subject. You could do it on your own, without a decade of schooling, but only if you're able to read whole books, and learn from those books what other books to read. The sort of "questioning" that is involved isn't like an "ask slashdot" where you're asking another person to tell you the answer, instead that sort of questioning involves reading lots of books, asking yourself a bunch of questions that come up, and then reading a bunch more books for the answer.
I think it is funny that you interpret claiming that a learning process involves work with being "misanthropic." Presumably, only haters are honest, and Good People blow scented smoke.
I'd really like to see the top 25 real estate cities for this stuff. Obviously not San Jose. SF, maybe; it wouldn't be the oddest thing in Golden Gate Park.
Right, everybody has to start somewhere, and the two sentences of my comment were too hard for you.
Everybody starts somewhere, but not ever subject is so easy that you can learn it by asking stupid questions. Some subjects require actual study, and if you're doing it on your own that means reading. Actual reading, too, not just chatting with people who know the answer.
If an intelligent person is asking "the most basic questions" in a way that might lead to a breakthrough, newsflash, they read the fucking manual before they asked any of those questions and they're asking the questions because they found a shortcoming in the manual. It does not mean they just stumbled up to the problem. If they're asking basic questions that they don't know the answer to, it means they're chatting with you. That's all it means.
It only took them a few months to convince me to turn off "Amber Alerts." If an alert goes out or not depends entirely on if the family has direct access to the police chief, it has nothing to do with the utility in the particular case of informing the public.
I'd love a "real emergencies only, no false positives, false negatives accepted" weather radio, though. No storm watches, and put the ocean weather on a separate channel. I really don't care what type of small craft advisory has been issued. I also don't need seasonal advisories about the first warm or cold day or the first fall rain storm.
Well, sure, if you don't understand what topics are included in a post, it will look wrong to you in various ways. But it would be more effective to understand what you're replying to, so your reply is on-topic.
Most of what you say is just some version of reality, twisted and spun really hard, the details forgotten. Like, IBM POWER series. It is true that its history and original instruction set were RISC, an instruction set created a decade before the thing that was actually implemented. So the books generally list it as RISC. And it is certainly true that you write code for that style of big iron that uses it in that way, and get really high performance. But they suck as workstations. The only reason it is nice as a user is because if you had that sort of thing, they spent 10x more for it and so it truly is 5x faster even at a desktop workload. Especially the later ones that would be most relevant, since they had pipelining and microcode and native CISC features even including x86 instruction support. It is kind of hilarious if that is the example of RISC system.
Also, you say IME doesn't come turned off, except it does come turned off, because you don't have access to Intel's microcode. That is logic salad, a total fail. This vulnerability is from the operating system driver being able to write to the IME. The ability to do that, the user-facing part of the hardware, is controlled in the BIOS by the user and it comes disabled. The OS cannot talk to the IME chip. That the IME chip has other code on it is not relevant. If you want a modern CISC system, and you want the fast fancy computer that can play neckbeard games, all the hardware is actually going to be microcode. And they won't tell you the details. If you want an x86 instruction set, guess what? It will not exist in hardware. When you buy an x86 from any vendor you're buying a black box x86 emulator. The only way around that, the only way to know what is really going on, is if you're running RISC hardware where the instruction set is actually implemented using physical registers and logic units that match what is described in the datasheet. And you won't be playing fancy games.
You will not know what is running on the other parts of the CPU, so why try to single out the IME as if you could know what is running on it? We're talking about a black box emulator. You don't know what it is "really doing." You have to trust the hardware vendor if you want to use the product.
Yeah, if you can't comprehend what features are on an integrated circuit, then you don't know what it does and shouldn't talk about it.
And if you don't fucking know, then you also don't fucking know and you don't even have to worry about if it is an integrated circuit or a toaster.
OTOH, if you know about IME and what it does, then there would be no surprise to find it on CISC CPUs intended for both servers and workstations. You would not tent to find those features on a RISC CPU because they're not being designed for server or desktop operating systems. You're not going to have sysadmins who have to manage PEBCAK needing it at all on RISC systems, even when on the same network, because the RISC systems are going to be appliances.
The features aren't about tinfoil, they're actually about how stupid the average IT employee is and the fact that they can't be trusted with control of their own workstation. Just look at the idiots on slashdot; half of them are this stupid for a living, using a computer.
Stopped reading at the R-word. You really need to work on your "theory of mind," because what kind of person is actually going to read that sort of thing?
The ME have nothing to do with the microcode[blah blah blah]
Absolutely, never said differently. I was broadly addressing two separate idiocies, one relating to IME, the other relating to people whining about Intel's CPUs having extra secret code that they don't understand the reasons for. For most of the slashdot, that is just one blurry thing about Intel CPUs being like systemd.
As for auditing, you can't audit the hardware in a CISC system. End of story. There is no chance. You don't really even want a processor if need that, you want programmable logic.
That's way more tinfoil than I'm gonna install, that's for sure.
You might want to check the conduit where you bring your internet link through the cage, and make sure there are no gaps.
I do want to say though that if you go and volunteer at a local shelter you'll be able to verify what I said. There are generally not individual electronic records even created. And they wouldn't, because some of the people they're helping have even more tinfoil than you.
That's a bunch of nonsense, it comes turned off. You don't "actually remove" part of the CPU, and you don't know how quickly it would fail because you've never spent the time to xray the chip and find the part you want, and lase it out while the chip is running. And then repeat until you get the process working. Nobody would do that.
You can't get around the fact that if you want a CISC computer, which includes any computer with significant branch prediction and CPU cache (required for to run any sort of neckbeard games) then your CPU doesn't even implement the hardware that is exposed to the user. With a RISC CPU the registers are probably real hardware registers, actual analog switches. In a CISC CPU, who knows what the registers are, they're probably just memory locations. The analog switches are multiple levels of abstraction below what you can interact with. You can't give the programmer access to what is really going on, and also provide the features that programmers want. Everybody who actually does the work want the CPU to expose an interface that acts like it is real old-school hardware, but no CPU could actually be built that way and have high performance.
Just pointing at complexity and asserting that it is pedantry isn't very interesting. There are real reasons why CISC CPUs must use microcode, and why the programmer won't have access to that layer. If you don't like it, just stick to RISC CPUs and don't ask for fancy graphics, branch prediction, or high IO bandwidth. Also, don't use any built-in CPU peripherals like DMA or I2C support, you're going to have to bit-bang that stuff to avoid accidentally using a peripheral IP block running mystery code.
If you'd own your statement, you might say something. But all you did is derp on my screen.
Are you saying that features you can't control are features you don't want? That has a different meaning than the stupid shit you actually said.
And in this case, if you didn't want microcode that you can't control you wouldn't buy a CISC computer at all. If you want to own a computer that can play graphical games, for example, you're already asking for microcode. You might not have the technical knowledge to understand that, but that doesn't change what is true. Nobody is going to build a CISC computer that gives the user control over the microcode because that would add a huge amount of engineering and QA. It is way better to just make it so that it mostly can't be updated.
Of course, the features in the story, that's all stuff that you can control. That's the whole point; you can control it, so if you accidentally install "malware," (or your OS does it for you) then it will have access to certain features! If you had those features turned off then you're fine. That's the situation here; if you turn IME on in your BIOS, and then install malware on Windoze, that malware can use the windows update features to patch an IME driver in memory, and thereby get access to it.
It also says right in the story that it is turned off by default. So clearly you do control it.
I'm not saying all Off Grid Guys, I'm just saying that at least one of them is actually living it.
If you think a drone flying over you means you're being monitored by the NSA, you might as well just crawl back into the tinfoil coffin, it still isn't safe to come out.
People don't even understand the difference between CISC and RISC processors, so how can they be expected to understand an issue like this?
They can't, and so they'll be credulous of all sorts of insane nonsense about "backdoors" and "shit... no one was asking for." Except that, it is shit people were in fact asking for.
"Give us features to manage security at the network level, we have too many hosts to do it at the host level and we have security PEBCAK." And so they implement network level management features that can manage the security without the intervention of the idiot at the keyboard, and whiners cry "backdoor, backdoor" because they don't realize that it comes turned off by default.
"Please make our games run faster, we need more throughput and better branch prediction." And so they upgrade the microcode and whiners whine about "extra shit" because they don't understand that you can't build a fast CISC processor in analog, you have to actually build a big mishmash of multiple RISC processors that together simulate a single CISC processor. To make it easy enough to program, it appears as a single thing, it even has fake timing semantics that are phrased as if it is a traditional digital logic circuit built from analog parts, but those analog parts are actually multiple levels of code below the interface you can touch. This goes back to the 1950s! Fast computers on the general market were already using microcode in the early 1960s! If you don't want microcode, stick to running RISC processors, and don't ask for fancy games. You can have tight real-time timing, but you won't get high total throughput or a useful cache.
We even have ignorant stuff on slashdot about "a separate chip inside Intel CPUs" which doesn't even show knowledge of what a "chip" is. And there are numerous IP blocks inside of any processor, the implication that there is something extra there is rather lame. Even an 8 bit RISC microcontroller has numerous IP blocks!
That will still be their response, and you should be able to detect by it if they're eating crow at all, or if you're just daydreaming.
It may be that it is secure and it is a standard feature that many companies want, and that the people complaining are in fact not only dunces but clear ignoramuses.
You can turn the feature off. And it isn't an obscure feature; it is an enterprise feature. There is actually a difference.
Everybody who knew said all along that if you add enterprise-level management software, it becomes an attack vector. Nobody disagreed with that. It is also a security vector, and for the same reasons. Yes, having a more powerful management tool enhances the power of whoever is controlling it, and malware on the network does happen in practice. It isn't perfect.
But lack of perfection does nothing to show that there is a problem, or that they'd be more secure without the feature. Obviously individuals and small business are better off without it, but any workstation or business computer comes with those settings in the BIOS. The only computers without the settings are the cheapo consumer stuff that don't even have these features anyways.
Amish who use cell phones, yeah, they might be monitored. Usually though the Amish keep a land line in the barn for needed business. Land lines are not monitored in bulk.
Prisoners are obviously heavily monitored, but why would the NSA be doing it? NSA isn't a law enforcement agency, and very few prisoners would be of interest to them. Furthermore, most of the prisoners only have access to landline payphones.
Deaf people using TTY are a group with more than zero members. Here you just wave your hands and manufacture a new type of monitoring that is unlikely to exist, and certainly not known to you.
Staying in a shelter does not create an electronic footprint, and creating an electronic footprint is not the same thing as being tracked by the NSA, or by any sort of feds. Most shelters don't even have electronic records at that level, they have 3x5 cards that get date stamped in ink, and the daily totals of people served are then input into a computer. Periodic aggregates will be reported to various places, especially to donors, but they're sure not going to publish a list on the wall of the people who stayed there.
You seem to have a lot of trouble finding the line between "zero" and "not zero." But it is actually narrow and clear.
Right, if you take half a construct it isn't the same as the whole thing.
Half of what I said isn't the same as the whole thing.
Barring a reporter based on the content of their speech is not the same thing as "barring a reporter."
Normally they wouldn't give a reason in that situation, so you wouldn't be able to prove it. But then that secrecy and ambiguity also limits their ability to use the action as a bludgeon to control what people say. But when they come straight out and say why they did it, apparently because they want to maximize the impact of the action, well then that changes everything.
There are lots of things that the government isn't supposed to do, that they normally can get away with, because they have good lawyers. If the decision-maker ignores the lawyers, they might no longer be able to get away with doing those things. That need for ambiguity is a limiting factor that is important to making government work even though all the individuals in it are biased and incompetent. They don't want to do the right thing, they're not trying to do the right thing, but they have to pretend that they're doing the right thing. The result is often that their actions are nearby to the right thing, and it works out.
Slashdot Mirror
Right, when you take a highly specialized niche and just put "X" in, then it sounds a lot better.
You seem to be denying that there would be any sort of subject where "asking a question" in the literal sense would be a useful way to gain knowledge. But there are in fact subjects that have too much required knowledge for that to work; you have to actually study the subject. You could do it on your own, without a decade of schooling, but only if you're able to read whole books, and learn from those books what other books to read. The sort of "questioning" that is involved isn't like an "ask slashdot" where you're asking another person to tell you the answer, instead that sort of questioning involves reading lots of books, asking yourself a bunch of questions that come up, and then reading a bunch more books for the answer.
I think it is funny that you interpret claiming that a learning process involves work with being "misanthropic." Presumably, only haters are honest, and Good People blow scented smoke.
One of the most popular technology teachers on youtube sticks to mechanical technology.
https://www.youtube.com/watch?...
I'd really like to see the top 25 real estate cities for this stuff. Obviously not San Jose. SF, maybe; it wouldn't be the oddest thing in Golden Gate Park.
Right, everybody has to start somewhere, and the two sentences of my comment were too hard for you.
Everybody starts somewhere, but not ever subject is so easy that you can learn it by asking stupid questions. Some subjects require actual study, and if you're doing it on your own that means reading. Actual reading, too, not just chatting with people who know the answer.
If an intelligent person is asking "the most basic questions" in a way that might lead to a breakthrough, newsflash, they read the fucking manual before they asked any of those questions and they're asking the questions because they found a shortcoming in the manual. It does not mean they just stumbled up to the problem. If they're asking basic questions that they don't know the answer to, it means they're chatting with you. That's all it means.
Yeah, but that is all wrong.
If you have to ask, it is too hard.
If it isn't too hard, you read the manual and already know what you need to know.
It is probably better for your mental health not to tell that story, just let it go and forget it happened. Acquaintances are over-rated anyways.
I disagree entirely. That is barbaric. Just hang them and be done with it, jeeze.
Like Churchill may or may not have said, "If you have to kill a man, it costs nothing to be polite."
Give him a last meal. Let him say some last words. And then just pull the lever and be done.
It only took them a few months to convince me to turn off "Amber Alerts." If an alert goes out or not depends entirely on if the family has direct access to the police chief, it has nothing to do with the utility in the particular case of informing the public.
I'd love a "real emergencies only, no false positives, false negatives accepted" weather radio, though. No storm watches, and put the ocean weather on a separate channel. I really don't care what type of small craft advisory has been issued. I also don't need seasonal advisories about the first warm or cold day or the first fall rain storm.
Always wrap the board in tinfoil, it doesn't matter how long you age the wood.
Well, sure, if you don't understand what topics are included in a post, it will look wrong to you in various ways. But it would be more effective to understand what you're replying to, so your reply is on-topic.
Most of what you say is just some version of reality, twisted and spun really hard, the details forgotten. Like, IBM POWER series. It is true that its history and original instruction set were RISC, an instruction set created a decade before the thing that was actually implemented. So the books generally list it as RISC. And it is certainly true that you write code for that style of big iron that uses it in that way, and get really high performance. But they suck as workstations. The only reason it is nice as a user is because if you had that sort of thing, they spent 10x more for it and so it truly is 5x faster even at a desktop workload. Especially the later ones that would be most relevant, since they had pipelining and microcode and native CISC features even including x86 instruction support. It is kind of hilarious if that is the example of RISC system.
Also, you say IME doesn't come turned off, except it does come turned off, because you don't have access to Intel's microcode. That is logic salad, a total fail. This vulnerability is from the operating system driver being able to write to the IME. The ability to do that, the user-facing part of the hardware, is controlled in the BIOS by the user and it comes disabled. The OS cannot talk to the IME chip. That the IME chip has other code on it is not relevant. If you want a modern CISC system, and you want the fast fancy computer that can play neckbeard games, all the hardware is actually going to be microcode. And they won't tell you the details. If you want an x86 instruction set, guess what? It will not exist in hardware. When you buy an x86 from any vendor you're buying a black box x86 emulator. The only way around that, the only way to know what is really going on, is if you're running RISC hardware where the instruction set is actually implemented using physical registers and logic units that match what is described in the datasheet. And you won't be playing fancy games.
You will not know what is running on the other parts of the CPU, so why try to single out the IME as if you could know what is running on it? We're talking about a black box emulator. You don't know what it is "really doing." You have to trust the hardware vendor if you want to use the product.
Right, but nothing is "hidden." Fuckin' duh.
Yeah, if you can't comprehend what features are on an integrated circuit, then you don't know what it does and shouldn't talk about it.
And if you don't fucking know, then you also don't fucking know and you don't even have to worry about if it is an integrated circuit or a toaster.
OTOH, if you know about IME and what it does, then there would be no surprise to find it on CISC CPUs intended for both servers and workstations. You would not tent to find those features on a RISC CPU because they're not being designed for server or desktop operating systems. You're not going to have sysadmins who have to manage PEBCAK needing it at all on RISC systems, even when on the same network, because the RISC systems are going to be appliances.
The features aren't about tinfoil, they're actually about how stupid the average IT employee is and the fact that they can't be trusted with control of their own workstation. Just look at the idiots on slashdot; half of them are this stupid for a living, using a computer.
Stopped reading at the R-word. You really need to work on your "theory of mind," because what kind of person is actually going to read that sort of thing?
The ME have nothing to do with the microcode[blah blah blah]
Absolutely, never said differently. I was broadly addressing two separate idiocies, one relating to IME, the other relating to people whining about Intel's CPUs having extra secret code that they don't understand the reasons for. For most of the slashdot, that is just one blurry thing about Intel CPUs being like systemd.
As for auditing, you can't audit the hardware in a CISC system. End of story. There is no chance. You don't really even want a processor if need that, you want programmable logic.
That's way more tinfoil than I'm gonna install, that's for sure.
You might want to check the conduit where you bring your internet link through the cage, and make sure there are no gaps.
I do want to say though that if you go and volunteer at a local shelter you'll be able to verify what I said. There are generally not individual electronic records even created. And they wouldn't, because some of the people they're helping have even more tinfoil than you.
No known ransomware runs on my pet rock, either.
That's a bunch of nonsense, it comes turned off. You don't "actually remove" part of the CPU, and you don't know how quickly it would fail because you've never spent the time to xray the chip and find the part you want, and lase it out while the chip is running. And then repeat until you get the process working. Nobody would do that.
You can't get around the fact that if you want a CISC computer, which includes any computer with significant branch prediction and CPU cache (required for to run any sort of neckbeard games) then your CPU doesn't even implement the hardware that is exposed to the user. With a RISC CPU the registers are probably real hardware registers, actual analog switches. In a CISC CPU, who knows what the registers are, they're probably just memory locations. The analog switches are multiple levels of abstraction below what you can interact with. You can't give the programmer access to what is really going on, and also provide the features that programmers want. Everybody who actually does the work want the CPU to expose an interface that acts like it is real old-school hardware, but no CPU could actually be built that way and have high performance.
Just pointing at complexity and asserting that it is pedantry isn't very interesting. There are real reasons why CISC CPUs must use microcode, and why the programmer won't have access to that layer. If you don't like it, just stick to RISC CPUs and don't ask for fancy graphics, branch prediction, or high IO bandwidth. Also, don't use any built-in CPU peripherals like DMA or I2C support, you're going to have to bit-bang that stuff to avoid accidentally using a peripheral IP block running mystery code.
It's not a feature if you can't control it.
That's just 100% pure derp.
If you'd own your statement, you might say something. But all you did is derp on my screen.
Are you saying that features you can't control are features you don't want? That has a different meaning than the stupid shit you actually said.
And in this case, if you didn't want microcode that you can't control you wouldn't buy a CISC computer at all. If you want to own a computer that can play graphical games, for example, you're already asking for microcode. You might not have the technical knowledge to understand that, but that doesn't change what is true. Nobody is going to build a CISC computer that gives the user control over the microcode because that would add a huge amount of engineering and QA. It is way better to just make it so that it mostly can't be updated.
Of course, the features in the story, that's all stuff that you can control. That's the whole point; you can control it, so if you accidentally install "malware," (or your OS does it for you) then it will have access to certain features! If you had those features turned off then you're fine. That's the situation here; if you turn IME on in your BIOS, and then install malware on Windoze, that malware can use the windows update features to patch an IME driver in memory, and thereby get access to it.
It also says right in the story that it is turned off by default. So clearly you do control it.
I'm not saying all Off Grid Guys, I'm just saying that at least one of them is actually living it.
If you think a drone flying over you means you're being monitored by the NSA, you might as well just crawl back into the tinfoil coffin, it still isn't safe to come out.
People don't even understand the difference between CISC and RISC processors, so how can they be expected to understand an issue like this?
They can't, and so they'll be credulous of all sorts of insane nonsense about "backdoors" and "shit... no one was asking for." Except that, it is shit people were in fact asking for.
"Give us features to manage security at the network level, we have too many hosts to do it at the host level and we have security PEBCAK."
And so they implement network level management features that can manage the security without the intervention of the idiot at the keyboard, and whiners cry "backdoor, backdoor" because they don't realize that it comes turned off by default.
"Please make our games run faster, we need more throughput and better branch prediction."
And so they upgrade the microcode and whiners whine about "extra shit" because they don't understand that you can't build a fast CISC processor in analog, you have to actually build a big mishmash of multiple RISC processors that together simulate a single CISC processor. To make it easy enough to program, it appears as a single thing, it even has fake timing semantics that are phrased as if it is a traditional digital logic circuit built from analog parts, but those analog parts are actually multiple levels of code below the interface you can touch. This goes back to the 1950s! Fast computers on the general market were already using microcode in the early 1960s! If you don't want microcode, stick to running RISC processors, and don't ask for fancy games. You can have tight real-time timing, but you won't get high total throughput or a useful cache.
We even have ignorant stuff on slashdot about "a separate chip inside Intel CPUs" which doesn't even show knowledge of what a "chip" is. And there are numerous IP blocks inside of any processor, the implication that there is something extra there is rather lame. Even an 8 bit RISC microcontroller has numerous IP blocks!
I thought they said it was 100% secure...
I don't believe either half of that; you didn't really think about it, and they didn't actually say that.
That will still be their response, and you should be able to detect by it if they're eating crow at all, or if you're just daydreaming.
It may be that it is secure and it is a standard feature that many companies want, and that the people complaining are in fact not only dunces but clear ignoramuses.
You can turn the feature off. And it isn't an obscure feature; it is an enterprise feature. There is actually a difference.
Everybody who knew said all along that if you add enterprise-level management software, it becomes an attack vector. Nobody disagreed with that. It is also a security vector, and for the same reasons. Yes, having a more powerful management tool enhances the power of whoever is controlling it, and malware on the network does happen in practice. It isn't perfect.
But lack of perfection does nothing to show that there is a problem, or that they'd be more secure without the feature. Obviously individuals and small business are better off without it, but any workstation or business computer comes with those settings in the BIOS. The only computers without the settings are the cheapo consumer stuff that don't even have these features anyways.
Amish who use cell phones, yeah, they might be monitored. Usually though the Amish keep a land line in the barn for needed business. Land lines are not monitored in bulk.
Prisoners are obviously heavily monitored, but why would the NSA be doing it? NSA isn't a law enforcement agency, and very few prisoners would be of interest to them. Furthermore, most of the prisoners only have access to landline payphones.
Deaf people using TTY are a group with more than zero members. Here you just wave your hands and manufacture a new type of monitoring that is unlikely to exist, and certainly not known to you.
Staying in a shelter does not create an electronic footprint, and creating an electronic footprint is not the same thing as being tracked by the NSA, or by any sort of feds. Most shelters don't even have electronic records at that level, they have 3x5 cards that get date stamped in ink, and the daily totals of people served are then input into a computer. Periodic aggregates will be reported to various places, especially to donors, but they're sure not going to publish a list on the wall of the people who stayed there.
You seem to have a lot of trouble finding the line between "zero" and "not zero." But it is actually narrow and clear.
Off Grid Guy.
Many of the people in prison.
Amish people.
Deaf people who still only use the old TTY system.
A few of the homeless.
Not everybody even has a cell phone or internet access, so it is unlikely to be everybody.
Grampy, that was part of the Cold War.
You don't get to argue with me when you're agreeing with me.
Right, if you take half a construct it isn't the same as the whole thing.
Half of what I said isn't the same as the whole thing.
Barring a reporter based on the content of their speech is not the same thing as "barring a reporter."
Normally they wouldn't give a reason in that situation, so you wouldn't be able to prove it. But then that secrecy and ambiguity also limits their ability to use the action as a bludgeon to control what people say. But when they come straight out and say why they did it, apparently because they want to maximize the impact of the action, well then that changes everything.
There are lots of things that the government isn't supposed to do, that they normally can get away with, because they have good lawyers. If the decision-maker ignores the lawyers, they might no longer be able to get away with doing those things. That need for ambiguity is a limiting factor that is important to making government work even though all the individuals in it are biased and incompetent. They don't want to do the right thing, they're not trying to do the right thing, but they have to pretend that they're doing the right thing. The result is often that their actions are nearby to the right thing, and it works out.