If students had to get 10/10 on each question in order to pass the class (by analogy with your rent payment) then there would be no distinction between passing (10/10) and excellence (also 10/10).
I'm sorry to hear that. As an instructor, I always encouraged my students to show their work for an entirely different reason.
As I would explain: if you do the problem correctly, you will get full credit. If you get the problem wrong, I will go through the work you've shown and try to give you as much partial credit as I can justify. If you don't show much work, I can't give you any partial credit and so you'll get zero points on the question.
This is the only fair way to do it. Students that get 90% of a problem right should get 9/10 possible points. But to do that, you really do have to encourage them to show their work in sufficient granularity for the instructor to grade it.
Yeah, there's definitely no legitimate reason that an upcoming project would decide to deprecate support for an OS that will be 10 years old at the point of release (Win7 came out in '09). Supporting and doing quality assurance on multiple OS targets is totally free from an engineering and testing standpoint. All API features from newer OSes are backported to decades-old ones.
Note that LibreOffice dropped support for OSX 10.8 (2012) and required various Linux components (Kernel/GTK) from 2006.
Removing support for old stuff at the right time is part of the software flow. Supporting the everything-on-everything model means less resources (both development and testing) for other stuff. Surely there's a "too soon" for deprecation" but also a "too late". One decade sounds pretty dang reasonable.
You've got the relationship backwards. A larger Goliath was annoyed at a smaller Goliath and enlisted the right David, gave him a twenty-million-dollar-nuclear-enabled-custom-silicon-powered slingshot.
The fact that this slingshot originated with Xiph needs to be noted alongside the fact that the current AOM bears nearly no resemblance to it.
To me, the critical thing to grok here is that there were sufficient powerful interests that wanted MPEG dead and were willing to coordinate and to spend freely to make it happen. The fact that we are going to get native hardware support in the next generation of everything for AOM formats seals the deal and makes my point: you have to spend millions over long lead times to bake things into silicon.
This is all a Good Thing(TM) because in this case the AOM solution is preferable to the MPEG solution, but it's definitely not some David-beats-Goliath scenario where some kid in a garage takes down the big-bad using the magic of open standards. In fact, if anything, the forces behind AOM are (and proved to be) even more Goliath-y than MPEG and it shows that industry power has shifted in their direction.
So less David-beats-Goliath and more like a bigger Goliath decided not to put up with a mini-Goliath anymore. Or, if you prefer, the king is dead, long live the king.
I'm pretty sure they looked at the install numbers for the feature pack in the store and concluded that if no one else cares they aren't going to either.
Uh, if you use BitLocker, FileVault, LUKS or (True|Vera)Crypt or other disk encryption (full-disk or separate volume) then your memory must contain the master keys for any unlocked partitions.
There was one (never accepted) patch for Linux, TRESOR, that would actually keep the key in some borrowed X86_64 registers. This was intended as a mitigation against cold-boot attempts, but could be repurposed here.
Yes, there's no argument you can do a lot of things much cleaner with a bare-bones RTOS.
Then a few years pass and your boss needs to:
Update the UI to support a new screen size
Update the UI and HID to support a new touchscreen model
Add ADA mandated audio prompts to the headphone jack
Accept chip-card transactions as well as magstripe
Reject magstripe attempts for a card with a chip
Accept NFC enabled debit cards
Accept the new cash-counting accessory for people to make cash deposits
Accept the new check-deposit accessory for people to make check deposits
...
Then your bare-bones RTOS isn't looking so hot. Who knows what shit-tastic GUI library or HID parsing they wrote for it. Meanwhile your boss's boss's boss is wondering why the hell we can't update these things like everyone else can and the security folks are clamoring to get chip & PIN working while you are staring down who-knows-how-they-built-it pile of WTF.
I mean, stop for a second and think, there are reasons that we don't just hire 13 year old computer whiz kids to implement everythingÂin their favorite obscure OS. Business requirements are a real thing, and they are a moving target.
Of course, Embedded Linux is a perfectly good choice for an OS. Still need libraries/frameworks for GUI, Audio, HID and peripherals. And then figure out how you are going to take kernel security updates without breaking ALSA/PulseAudio, or else pay RH to do it for you. By the time you are done it's not going to be "slim" and it definitely won't be a a bare-bones RTOS.
Consciousness is the subjective experience of an analytic process that is sufficiently advanced to analyze itself (and to analyze itself analyzing itself, ad infinitum). Reflexiveness is the sine qua non of consciousness.
- The one where stuff becomes sufficiently complicated and (insert long and bamboozling set of random examples of complicated systems and a grab bag of barely coherent ideas) and presto consciousness happens for some reason, apparently (c/f strange loops etc)?
The one where life spontaneously forms, spends a few billions years mucking about (literally), the Cambrian Explosion happens creating more complex forms, gradually resulting in species with a system dedicated to gathering and analyzing sensory input. Since even for crude organisms, more analysis can sometimes lead to increased more fitness (not always, see crocodiles), it's not unreasonable that the end product was a species with some much analytic capability it could analyze itself, and analyze itself analyzing itself, and analyze itself analyzing its own strange love for late night comedy television.
Reminds me of a pithy saying: being a democratically legitimate leader isn't entirely about being elected, it's about stepping down when someone else gets elected. We don't owe respect or legitimacy to elected leaders whose principle is one man, one vote, one time.
The Western world has been taken too many times by this. Elections are a necessary but not sufficient condition for democratic government.
This attitude makes no sense. Sure it's "perfectly good hardware", but it costs real time and effort to keep updating* it. As time goes on and the PGH becomes rarer and rarer, the ratio of benefit from keeping support relative to spending that effort on something else becomes less and less favorable. Eventually, even if the hardware is perfectly good, it doesn't make sense to continue to support upgrading it because so few remain.
It's also incredibly rude to people that volunteer their support to build a free kernel to suggest they are "lazy" for not supporting your Pentium III. The whole of Linux is their gift to you, and you can continue running the existing kernel on it as long as you like, but instead you have to insult a gift horse in the mouth over it?
* Note that much of this support cost is probably QA. You can't just ship it because you'll end up like Microsoft where everyone was pissed at them recently for bricking some old AM2-based system that they claimed to support.
You are correct about the scientific fact that producing meat requires out of cycle CO2.
You are also correct that the OP was ignorant of this fact.
You are absolutely not correct that OP's ignorance makes the problem intractable. It's intractable for completely other reasons.
Chief among those reasons is that many (not all) people strongly prefer to eat meat. And given that (for better or worse) people have free will, they will enact that choice both individually through their dietary choices and collectively through policies that enable those choices.
I hope that people that are serious about climate change understand that, because I don't see us making much progress on it otherwise.
I think I understand your claim (not 100% agreeing or disagreeing).
What I don't understand is the difference between a solution that "requires doing something" versus "requires not doing something". As I see it, both are something that can be stopped later in time.
For instance, we might have (among others) a solution of "not burning coal". In 50 years, however, we might stop that solution by starting to burn coal again.
I guess my point is anything can be stopped, including stopping...
Shouldn't it then also be regardless of the 'what'?
I meant, irrespective of whatever (combination of) solutions we chose, it's possible that eventually we will stop. That's true in the most trivial sense, but what I don't is why it has impact on the merits of any particular solution.
... and, besides, they can no longer afford to pump aerosols into the atmosphere.
The cost of such a program, especially after it's been going for decades, is minuscule compared to the cost of carbon reduction. The idea that we'll suddenly not be able to afford it is nuts, but moreover, it's applicable a fortiori to any other plan. Who would claim that "well, we could cut carbon emissions, but then in 2050 we might no longer be able to afford it and go back to coal, which would be worse" is a legitimate argument against carbon reduction?
There are a million legitimate objections to geo-engineering. This one, however, is total nonsense.
That is a difference. I don't see how it's a key difference who holds the record.
In the bank/oil-co/bad-guy example, does using a 3rd party IT department instead of in-house change things? Or any number of intermediaries can be added: Oil company contracts ITCorpUS, ITCorpUS has a subsidiary in Ireland, Ireland has a subsidiary in Cook Islands....
I think it's a bit disingenuous to say "the guy that can touch the machine" is the one who makes final decisions. That's certainly not the case in all the companies I've worked for: the guy accepts what management tells him. If he doesn't, he will be replaced by someone that does. No organization (no matter how enlightened) gives the IT dudes the final authority over who gets access to what systems.
"Hey Bob, did you get a SCM account?" "No, I cut off the sysadmin for that system in the parking lot and my boss says he's the only one that can make an account. Tough luck I guess, can I borrow your login"
I would imagine that Microsoft USA has constructive control over the datacenter operations in one form or another. They might not have machine permissions in the narrow technical sense, but effectively they can
[ Reminds me a bit about the Planet Money about shell corporations. They would hire local dudes in the Cayman Islands or wherever to be shareholders and board members of a company. But they would then sign a separate contract with the local dude saying that the local dude will vote however some tax-dodger says they will.
The claim then is that the tax scofflaw doesn't "own" the shell corporation (technically true: he has no shares) and so it doesn't have to be reported on their taxes. Meanwhile, through the voting contract, they get to control the company assets to pay for shit they want. The IRS decided pretty decisively that this BS doesn't fly -- you are the beneficial owner of a corporation based on the actual facts, not the nominal status.
Of course, analogies are imperfect so I'm fine if you think this isn't entirely apt. But the parallels are a bit striking: Microsoft USA probably has constructive control over the datacenter, we'll see what the courts think soon I guess ]
Prosecutor: The defendant has not turned over emails between their executives discussion the probability of an (oil leak) (fiscal collapse) (other bad thing).
Judge: Why not? Defendant: Those emails are not stored in this country.
Judge: Which country are they stored in? Defendant: Please refer to the statement from EvilCO IT explaining that our emails are stored in a database that is then sharded across all our subsidiaries around the world.
Judge: And you need the pieces, the shards, from all the countries to reassemble them? Defendant: There's some redundancy for catastrophic failures, so no, not all.
Judge: But most. Defendant: Yes.
Judge: And this can be done from a server here in the United States. Defendant: No, email accounts are managed by those subsidiaries in those countries. Our EvilCO IT here doesn't have permissions to create them.
Judge: But when a new employee starts, they get access? How? Defendant: Yes. The US office requests an account for them.
Judge: So you do have permissions to create the new accounts Defendant: No, only the administrators at the subsidiaries do.
Judge: But you can tell them to do so. Defendant: Yes
Judge: And they can't say no. Defendant: I'm not aware of a sysadmin in a subsidiary refusing to create an account for their local shards to an employee authorized by corporate.
Judge: So you claim you don't have permission, but if you make a request then it's always fulfilled. Defendant: It's fulfilled by an administrator in the subsidiary that has permission.
Judge (daydreaming): Bailiff, please tase this lawyer in the balls repeatedly until he stop this bullshit. Judge (IRL): Counsel, I think you have constructive access if every time you request access, someone with actual permission grants it. Defendant: Multiple administrators are required, in each subsidiary, to grant access to the shards necessary to access the email system.
Judge: Yes, multiple, OK, it's a nice shell game. How about turn over the documents about your damned oil spill already?
Is there anybody left at this point who actually believes that the US intelligence apparatus is governed by laws?
If that's the case, why would they bother renewing the statutory authorization?
And why would we bother fighting the renewal of the statutory authorization?
The very fact that we are all debating the passage of a law seems to be a fairly strong indicator that the US intelligence apparatus is at least partially beholden to the authorization (or lack thereof) provided by the legislature.
Only anglophones have the right to primary english education for their children. If francophone parents wish to send their kids to the english school, they are shit outta luck.
The following children, at the request of one of their parents, may receive instruction in English:
1) a child whose father or mother is a Canadian citizen and received elementary instruction in English in Canada, provided that that instruction constitutes the major part of the elementary instruction he or she received in Canada;
2) a child whose father or mother is a Canadian citizen and who has received or is receiving elementary or secondary instruction in English in Canada, and the brothers and sisters of that child, provided that that instruction constitutes the major part of the elementary or secondary instruction received by the child in Canada;
Itâ(TM)s also true that the population of developers with interest and aptitude in a given piece of software will shrink the higher up the stack you go.
Really? More people have aptitude for writing bare-metal firmware than some cloud application? I doubt it, not so much due to "difficulty", but due to the fact that there are far more jobs at lofty levels of abstraction where you can do useful work than there are down in the kernel/metal.
Slashdot Mirror
If students had to get 10/10 on each question in order to pass the class (by analogy with your rent payment) then there would be no distinction between passing (10/10) and excellence (also 10/10).
Dude, if you do it the fast way and get the right answer, you'll get 100% credit in my book.
But if you do it the fast way and make a tiny mistake and write nothing down but the wrong answer, how can I give you any more than 0%?
I'm sorry to hear that. As an instructor, I always encouraged my students to show their work for an entirely different reason.
As I would explain: if you do the problem correctly, you will get full credit. If you get the problem wrong, I will go through the work you've shown and try to give you as much partial credit as I can justify. If you don't show much work, I can't give you any partial credit and so you'll get zero points on the question.
This is the only fair way to do it. Students that get 90% of a problem right should get 9/10 possible points. But to do that, you really do have to encourage them to show their work in sufficient granularity for the instructor to grade it.
Yeah, there's definitely no legitimate reason that an upcoming project would decide to deprecate support for an OS that will be 10 years old at the point of release (Win7 came out in '09). Supporting and doing quality assurance on multiple OS targets is totally free from an engineering and testing standpoint. All API features from newer OSes are backported to decades-old ones.
Note that LibreOffice dropped support for OSX 10.8 (2012) and required various Linux components (Kernel/GTK) from 2006.
Removing support for old stuff at the right time is part of the software flow. Supporting the everything-on-everything model means less resources (both development and testing) for other stuff. Surely there's a "too soon" for deprecation" but also a "too late". One decade sounds pretty dang reasonable.
You've got the relationship backwards. A larger Goliath was annoyed at a smaller Goliath and enlisted the right David, gave him a twenty-million-dollar-nuclear-enabled-custom-silicon-powered slingshot.
The fact that this slingshot originated with Xiph needs to be noted alongside the fact that the current AOM bears nearly no resemblance to it.
To me, the critical thing to grok here is that there were sufficient powerful interests that wanted MPEG dead and were willing to coordinate and to spend freely to make it happen. The fact that we are going to get native hardware support in the next generation of everything for AOM formats seals the deal and makes my point: you have to spend millions over long lead times to bake things into silicon.
This is all a Good Thing(TM) because in this case the AOM solution is preferable to the MPEG solution, but it's definitely not some David-beats-Goliath scenario where some kid in a garage takes down the big-bad using the magic of open standards. In fact, if anything, the forces behind AOM are (and proved to be) even more Goliath-y than MPEG and it shows that industry power has shifted in their direction.
So less David-beats-Goliath and more like a bigger Goliath decided not to put up with a mini-Goliath anymore. Or, if you prefer, the king is dead, long live the king.
I'm pretty sure they looked at the install numbers for the feature pack in the store and concluded that if no one else cares they aren't going to either.
Would you?
Uh, if you use BitLocker, FileVault, LUKS or (True|Vera)Crypt or other disk encryption (full-disk or separate volume) then your memory must contain the master keys for any unlocked partitions.
There was one (never accepted) patch for Linux, TRESOR, that would actually keep the key in some borrowed X86_64 registers. This was intended as a mitigation against cold-boot attempts, but could be repurposed here.
Yes, there's no argument you can do a lot of things much cleaner with a bare-bones RTOS.
Then a few years pass and your boss needs to:
Then your bare-bones RTOS isn't looking so hot. Who knows what shit-tastic GUI library or HID parsing they wrote for it. Meanwhile your boss's boss's boss is wondering why the hell we can't update these things like everyone else can and the security folks are clamoring to get chip & PIN working while you are staring down who-knows-how-they-built-it pile of WTF.
I mean, stop for a second and think, there are reasons that we don't just hire 13 year old computer whiz kids to implement everythingÂin their favorite obscure OS. Business requirements are a real thing, and they are a moving target.
Of course, Embedded Linux is a perfectly good choice for an OS. Still need libraries/frameworks for GUI, Audio, HID and peripherals. And then figure out how you are going to take kernel security updates without breaking ALSA/PulseAudio, or else pay RH to do it for you. By the time you are done it's not going to be "slim" and it definitely won't be a a bare-bones RTOS.
Consciousness is the subjective experience of an analytic process that is sufficiently advanced to analyze itself (and to analyze itself analyzing itself, ad infinitum). Reflexiveness is the sine qua non of consciousness.
- The one where stuff becomes sufficiently complicated and (insert long and bamboozling set of random examples of complicated systems and a grab bag of barely coherent ideas) and presto consciousness happens for some reason, apparently (c/f strange loops etc)?
The one where life spontaneously forms, spends a few billions years mucking about (literally), the Cambrian Explosion happens creating more complex forms, gradually resulting in species with a system dedicated to gathering and analyzing sensory input. Since even for crude organisms, more analysis can sometimes lead to increased more fitness (not always, see crocodiles), it's not unreasonable that the end product was a species with some much analytic capability it could analyze itself, and analyze itself analyzing itself, and analyze itself analyzing its own strange love for late night comedy television.
Seriously, are we out of real scientific problems to study?
Reminds me of a pithy saying: being a democratically legitimate leader isn't entirely about being elected, it's about stepping down when someone else gets elected. We don't owe respect or legitimacy to elected leaders whose principle is one man, one vote, one time.
The Western world has been taken too many times by this. Elections are a necessary but not sufficient condition for democratic government.
This attitude makes no sense. Sure it's "perfectly good hardware", but it costs real time and effort to keep updating* it. As time goes on and the PGH becomes rarer and rarer, the ratio of benefit from keeping support relative to spending that effort on something else becomes less and less favorable. Eventually, even if the hardware is perfectly good, it doesn't make sense to continue to support upgrading it because so few remain.
It's also incredibly rude to people that volunteer their support to build a free kernel to suggest they are "lazy" for not supporting your Pentium III. The whole of Linux is their gift to you, and you can continue running the existing kernel on it as long as you like, but instead you have to insult a gift horse in the mouth over it?
* Note that much of this support cost is probably QA. You can't just ship it because you'll end up like Microsoft where everyone was pissed at them recently for bricking some old AM2-based system that they claimed to support.
You are correct about the scientific fact that producing meat requires out of cycle CO2.
You are also correct that the OP was ignorant of this fact.
You are absolutely not correct that OP's ignorance makes the problem intractable. It's intractable for completely other reasons.
Chief among those reasons is that many (not all) people strongly prefer to eat meat. And given that (for better or worse) people have free will, they will enact that choice both individually through their dietary choices and collectively through policies that enable those choices.
I hope that people that are serious about climate change understand that, because I don't see us making much progress on it otherwise.
I think I understand your claim (not 100% agreeing or disagreeing).
What I don't understand is the difference between a solution that "requires doing something" versus "requires not doing something". As I see it, both are something that can be stopped later in time.
For instance, we might have (among others) a solution of "not burning coal". In 50 years, however, we might stop that solution by starting to burn coal again.
I guess my point is anything can be stopped, including stopping ...
Shouldn't it then also be regardless of the 'what'?
I meant, irrespective of whatever (combination of) solutions we chose, it's possible that eventually we will stop. That's true in the most trivial sense, but what I don't is why it has impact on the merits of any particular solution.
... and, besides, they can no longer afford to pump aerosols into the atmosphere.
The cost of such a program, especially after it's been going for decades, is minuscule compared to the cost of carbon reduction. The idea that we'll suddenly not be able to afford it is nuts, but moreover, it's applicable a fortiori to any other plan. Who would claim that "well, we could cut carbon emissions, but then in 2050 we might no longer be able to afford it and go back to coal, which would be worse" is a legitimate argument against carbon reduction?
There are a million legitimate objections to geo-engineering. This one, however, is total nonsense.
That is a difference. I don't see how it's a key difference who holds the record.
In the bank/oil-co/bad-guy example, does using a 3rd party IT department instead of in-house change things? Or any number of intermediaries can be added: Oil company contracts ITCorpUS, ITCorpUS has a subsidiary in Ireland, Ireland has a subsidiary in Cook Islands ....
I think it's a bit disingenuous to say "the guy that can touch the machine" is the one who makes final decisions. That's certainly not the case in all the companies I've worked for: the guy accepts what management tells him. If he doesn't, he will be replaced by someone that does. No organization (no matter how enlightened) gives the IT dudes the final authority over who gets access to what systems.
"Hey Bob, did you get a SCM account?"
"No, I cut off the sysadmin for that system in the parking lot and my boss says he's the only one that can make an account. Tough luck I guess, can I borrow your login"
I would imagine that Microsoft USA has constructive control over the datacenter operations in one form or another. They might not have machine permissions in the narrow technical sense, but effectively they can
[ Reminds me a bit about the Planet Money about shell corporations. They would hire local dudes in the Cayman Islands or wherever to be shareholders and board members of a company. But they would then sign a separate contract with the local dude saying that the local dude will vote however some tax-dodger says they will.
The claim then is that the tax scofflaw doesn't "own" the shell corporation (technically true: he has no shares) and so it doesn't have to be reported on their taxes. Meanwhile, through the voting contract, they get to control the company assets to pay for shit they want. The IRS decided pretty decisively that this BS doesn't fly -- you are the beneficial owner of a corporation based on the actual facts, not the nominal status.
Of course, analogies are imperfect so I'm fine if you think this isn't entirely apt. But the parallels are a bit striking: Microsoft USA probably has constructive control over the datacenter, we'll see what the courts think soon I guess ]
Prosecutor: The defendant has not turned over emails between their executives discussion the probability of an (oil leak) (fiscal collapse) (other bad thing).
Judge: Why not?
Defendant: Those emails are not stored in this country.
Judge: Which country are they stored in?
Defendant: Please refer to the statement from EvilCO IT explaining that our emails are stored in a database that is then sharded across all our subsidiaries around the world.
Judge: And you need the pieces, the shards, from all the countries to reassemble them?
Defendant: There's some redundancy for catastrophic failures, so no, not all.
Judge: But most.
Defendant: Yes.
Judge: And this can be done from a server here in the United States.
Defendant: No, email accounts are managed by those subsidiaries in those countries. Our EvilCO IT here doesn't have permissions to create them.
Judge: But when a new employee starts, they get access? How?
Defendant: Yes. The US office requests an account for them.
Judge: So you do have permissions to create the new accounts
Defendant: No, only the administrators at the subsidiaries do.
Judge: But you can tell them to do so.
Defendant: Yes
Judge: And they can't say no.
Defendant: I'm not aware of a sysadmin in a subsidiary refusing to create an account for their local shards to an employee authorized by corporate.
Judge: So you claim you don't have permission, but if you make a request then it's always fulfilled.
Defendant: It's fulfilled by an administrator in the subsidiary that has permission.
Judge (daydreaming): Bailiff, please tase this lawyer in the balls repeatedly until he stop this bullshit.
Judge (IRL): Counsel, I think you have constructive access if every time you request access, someone with actual permission grants it.
Defendant: Multiple administrators are required, in each subsidiary, to grant access to the shards necessary to access the email system.
Judge: Yes, multiple, OK, it's a nice shell game. How about turn over the documents about your damned oil spill already?
Is there anybody left at this point who actually believes that the US intelligence apparatus is governed by laws?
If that's the case, why would they bother renewing the statutory authorization?
And why would we bother fighting the renewal of the statutory authorization?
The very fact that we are all debating the passage of a law seems to be a fairly strong indicator that the US intelligence apparatus is at least partially beholden to the authorization (or lack thereof) provided by the legislature.
Only anglophones have the right to primary english education for their children. If francophone parents wish to send their kids to the english school, they are shit outta luck.
The following children, at the request of one of their parents, may receive instruction in English:
1) a child whose father or mother is a Canadian citizen and received elementary instruction in English in Canada, provided that that instruction constitutes the major part of the elementary instruction he or she received in Canada;
2) a child whose father or mother is a Canadian citizen and who has received or is receiving elementary or secondary instruction in English in Canada, and the brothers and sisters of that child, provided that that instruction constitutes the major part of the elementary or secondary instruction received by the child in Canada;
3) a child whose father and mother are not Canadian citizens, but whose father or mother received elementary instruction in English in Québec, provided that that instruction constitutes the major part of the elementary instruction he or she received in Québec;
4) a child who, in their last year in school in Québec before 26 August 1977, was receiving instruction in English in a public kindergarten class or in an elementary or secondary school, and the brothers and sisters of that child;
5) a child whose father or mother was residing in Québec on 26 August 1977 and had received elementary instruction in English outside Québec, provided that that instruction constitutes the major part of the elementary instruction he or she received outside Québec.
I'm really not sure either way, but could you provide a citation to the fact that LHR (for instance) is gate-constrained and not slot-constrained?
My initial thought was that it must be takeoff/landing slot-constrained, but now I'm not convinced.
Itâ(TM)s also true that the population of developers with interest and aptitude in a given piece of software will shrink the higher up the stack you go.
Really? More people have aptitude for writing bare-metal firmware than some cloud application? I doubt it, not so much due to "difficulty", but due to the fact that there are far more jobs at lofty levels of abstraction where you can do useful work than there are down in the kernel/metal.