I have little doubt that Amazon had the necessary corpus to generate books including ones containing all the elements that people look for in a good book. That said, well designed and programmed robots can make reasonable food out of correct quantities of each ingredients in the correct order, but after a while it will all just begin to taste the same and be bland. In the case of the software generated stories, some of it might be accidentally good enough to make it into a literary publications if all of the vogue themes show up; however, until such time that we have "electric readers" that will read for us so that we won't have to do so, I doubt the bulk of it will make the upper ranks for the bulk of readers, at least for the first few generations of such.
Copyright and patents are still useful, but the timeframe is just purely rent-seeking. How about $10 dollars to register either for 2 years, double the previous for each 1 year extension. $10, $20, $40, $80, etc. The first few years will be cheap and a common citizen could get a foothold. Someone successful could even continue the monopoly for a few extra years. After a while, even a company such as Apple with unruly amounts of capital would give up and allow it to be public domain.
I agree for the trademarks, but would advice that it only applies to the context of use as a trademark. Trademarking a certain mouse would not prevent the character of that mouse from being used in a fictional work if not meant to represent the company.
I am not so concerned with the presence of copyright since it would be one of the few things keeping my modest work from simply being ripped off by someone with more marketing skills than I before I can even attempt to do something. I am, however, concerned with the blatant abuse. If someone is using one of my random utilities or hanging a print of one of my ugly renders on the wall 20 years from now, I would have had a chance to market it. 100 years of rent-seeking, it only encourages people to rest upon their laurels while speeding up inflation by creating undue artificial scarcity.
During a trip earlier this year to Toronto, I had visited Fort York. At least one sign referenced "Americans" plundering and burning the place. Since I doubt they would being doing it to themselves, it stands to reason they were referencing the US soldiers (history class, years ago, touched upon it as an US offensive towards the British which, while successful, wasn't quite as strategic as hoped, the museum at the site of the fort suggests the same but in less sugarcoated language, for fairly plain reasons).
Canadians and Mexicans are indeed citizens of the North American continent (or some of the citizens of the larger American continent, depending on whose map you read), but "United Statians" does not roll off the tongue as well in English, so I imagine that why it became a de facto convention. Since it is not de jure, maybe it will change, maybe it won't.
Honey-pots are not the newest idea out there; however, I might point out that actual dead-code in the binary is often removed by compilers when using higher numbers on the optimization parameter. -O1 might not remove it, but -O3 probably would (compiler dependent). I might also make note that if there is room for a buffer overflow in a library, the inside of the if statement might make an attractive jump target if is it visible to the context (probably involving setting EAX, ES, and EDI first and doing a direct far jump to the address with the current stack).
An interesting idea, but cost and risk versus benefit, I do not think it worth it.
People are using their phones increasingly more like a low-end laptop. I suspect the bargain cellphones will eventually catch up, but in the meantime, my $400 will not do the stuff that I would like to do. While I am more willing to forego doing the extra things, some people have more dollars and less restraint. We share see what the market will bear once the data is bare.
Maybe that Perl of humor was the straw that broke the camel's back. That said, at the very least, Perl is still easier to read than Whitespace and easier to pronounce than INTERCAL.
I note that I feel the same "twitch" in the ears as I do just after using a firearm. I understand there is a muscle there; perhaps it is learned reflex to the visual cue.
Cause for many of us Gen-X'ers, investments haven't even broken even. Heck, we're going to be the first generation to pay more into Social Security than we'll get back.
That about sums up investing for Gen-X'ers.
I'm on the line between Gen-X'er and Millennial. After considering taxes and fees, my mutual fund near about breaks even: if sold, there would be just about enough left over to buy a grande latte and leave a decent tip.:-P
I do not think that given his potential income that "coffee shop" coffee will make or break him. On the other hand, someone that that makes far less could put that same money into savings to get a bit shy of an extra thousand a year of financial buffer, assuming that they do not choose to divert it elsewhere. Of course, the few minutes of time making his own coffee could be invested elsewhere, but if anything like my significant other, then that task is likely being "worked-in" as part of multitasking another close proximity task such as making breakfast. Before I started my career, I usually eschewed "barista made coffee" except when "treating myself" after passing my finals or getting high enough performance reviews to qualify for a bonus. Now since I make a reasonable salary, I am a bit more relaxed about "extras." That said, as per the adage, "if you want something done right, do it yourself": I like my coffee a particular way, so I still tend to brew it myself. Also, the gentleman, who probably does it purely on principle, is free to make his own choice to brew it himself just as the person that opts to pay a premium to have it done for them.
Once logged in, it takes about 10 minutes on the average Linux machine and about 15 on the average Windows 7 with Cygwin machine, including coffee breaks, to setup a private certificate authority, convert the public key cert, and copy it to a flash drive. Installing as a trusted certificate authority onto multiple Windows machines can be done rapidly using group policy, if available, or Powershell script from an admin account if not. It takes a couple of minutes per CSR to sign for the webservers and random numbers of minutes to get them into IIS. Apache is a bit faster.
That said, I have not read the details of the change, but I hope there is a readily accessible power-user option to turn it off. I worked for several companies with IT department that demands exclusive control over whatever CAs are out there but takes forever fulfilling new signing requests. I am not at one now, but through the lens of past experience, it is better to have the option to control the behaviour.
As stated, seems useful for investigation/obtaining warrant, and accuracy can be confirmed with blackbox quality assurance. On the other hand, I would refrain from using as "star evidence." That said, if I were in there shoes, I would get a my lawyers to draft up a nice pair of NDAs, get a respected university to verify the science, and get a security company to review the coding to get a pair of gold stars. It might make DA's and investigators a bit more likely to take a look.
You can scale an image to a fixed size, divide it up into parts, take a histogram of each part (Fourier transform), and shove it into an indexed database (might need to duplicate also do greyscale separately). If another image has most of the parts' histograms within a certain arbitrary percentage, the picture is probable that it is the same (no need to directly compare the original images). That, in the loosest way possible, can be considered a "hash comparison." Of course, I'm pretty sure Facebook has something more advanced than that bit of minor experimentation. Once confirmed not to be the same as a previous image, just do the normal built-in facial recognition. If unable to automatically confirm, send it to queue for an underpaid human to handle.
My "photo" is a vector art drawing of myself. Since it would have been no fun tracing a picture of myself in Inkscape, I opted for kinda-sorta manga style. I doubt a non-human can make a comparison (I can make a face humanoid and maybe even identifiable, but art isn't my forte). More than likely automatic comparison is done when it identifies photographs. For everything else, the upload likely generates a ticket for a human to pull up and click 'Yes', 'No', or 'Maybe' with 'Yes' making the system happy, 'No' suspending the account with some message, and 'Maybe' to assign the ticket to someone else or pop up a call customer-service message.
My experience is that readability depends upon the quality of the code itself. Python, for example, while usually lending itself to being readable can be very difficult to read when poor choices are made with coding standards or at the very least they are not followed. This is particularly true when using the "more advanced features."
In regards to your sample:
while(*d++=*s++);
If no reason was given, I would quickly demand:
strcpy(s,d);/* include string.h above and do not use with std::string */
Unless optimized inline, there would be extra work on the stack for the call, but someone unfamiliar with the idiom or the subtleties of operator precedence (or C lacking a dedicated bool type) of the original sample will have trouble. Yes, a veteran C/C++ programmer would recognize it instantly, but for crunch-time, it is not that unusual for a "loaner" to be attached to the project that normally uses a different but similar language, e.g., a skilled Java or C# person that knows what a pointer is. The latter example I provided has a clear purpose. Of course, extra help worth their salt would have the sense to look it up, assuming they aren't on their last drop of liquid motivation while burning the midnight oil.
This is related back to your mention of:
The key isn't in avoiding symbols, which you're supposed to be able to just read if you plan to do any coding at all, but to ensure formatting and sane use of whitespace makes turns a string of symbols into a set of instructions.
I am just adding that it helps to be mindful of having good naming, try to use standard libraries (except for when there is documented good reason not to do so), document when doing something esoteric (when cannot be helped), and keep the "clever" to the obfuscation contests and proof-of-concept snippets.
For the TL;DR, having good conventions (whitespace, naming, library usage, etc), inline documentations (and actual written documentation, please), reasonable communication concerning standards, and clear expectations can allow for C/C++, or most other languages for that matter, to be maintainable even in the largest of projects.
An extra after thought, the effect of a JIT compiled language may be possible at the kernel level by "faking it." While there would be limitations, upon driver installation, native image generation could be done. The kernel would need to have some sort of metadata service. Garbage collection might could be a compiler service as well, but I would opt for having the initial code compile do static analysis for reference count and just inject deallocations after the last reference. Any case that cannot be verified to remain in scope would fail the compile; the programmer would have to revert to manual allocation/deallocation possibly disguised as support libraries and use at their own risk. Additional keyword or syntactic sugar might could be used when the programmer need to delay deallocation for some reason such as performance. Use try/finally (or equivalents such as try-with-resources, using keyword) for setting allocation boundaries or accessing scarce resources. Use case would be for rapid development of drivers for cross-architecture but same kernel where the driver does not require much direct memory access or IO, but would benefit from not having a context switch.
I doubt firmware would be possible any time soon with JIT compile languages. Considering CPU speeds and memory sizes, garbage collection should be possible in firmware (after some bootstrap); however, it would be quite the waste considering the target. I would say similar for kernel drivers. User-land drivers such as certain types of human interface devices/filters, printer drivers, scanner drivers, maybe (big maybe) virtual file system drivers, or similar which would delegate its communication with hardware through an abstraction layer would at least be worth considering on systems with sufficient resources to make it worth while. Although, at that point, I would sooner call those services.
I was playing the game for a while (until schedule and loss of interest took the fun out of it). I never walked into the road without looking even once; I will admit to colliding with other players of the same game a few times in a park area when not using the augmented reality feature due to battery-life and ease of playing. On the other hand, while driving, I have had several people walk in front of me while playing (the ball toss "finger-flick" is quite distinctive) without so much as a care in the world; however, I also had similar numbers of people with both thumbs on the phone, apparently composing something, suddenly walk in front of me. My reaction time is not excellent, so if driving conditions were different, there may very have been different outcomes than me stopping and tapping my horn: we (in each case) both were lucky. While the majority were embarrassed and/or apologetic and quickly removed themselves from the right-of-way, one particular case went so far as to showing the center-finger-salute and then repeatedly hitting my vehicle while continuing to yell about it being my fault for them not successfully capturing a Bellspout, at least until I drove off. Fortunately, I never had any incident concerning a younger child running into the right-of-way while apparently playing that game. I did see, however, a few (seemingly) responsible parents taking active measures to prevent disaster, e.g., calling back child before he/she arrives at road, placing themselves between child and road, or at least intercepting the child and extracting mobile device. I strongly hope that of the parents that have children playing the game, those samples are not outliers.
Pokemon Go (and Ingress for that matter), for those immersed in the game, is indeed a distraction in the same way social media and other electronic communications are. That said, I would hesitate to ban it or provide any unnecessary movement restrictions. I am prone to be easily distracted myself; however, I can choose to forgo playing a game, texting, etc for a few moments while it is unsafe to do so. That is my responsibility. I suspect that only a tiny minority of adolescents and adults have such significant developmental, functional, or cognitive conditions that would make it impossible (or at least exceedingly difficult) to train one's self to exercise caution. For those that would have such troubles (or anyone, just for doing so), if wanting to join in playing, go with a group of friends, have fun, and be safe.
While I certainly not found of C++ myself, I can certainly code in it. As it stands, most of the problems that I have to solve on a regular basis does not involve usage of that language directly. That said, I still think there is use for it, even beyond legacy compatibility. I typically work on user-land applications, data ETL jobs, and web services. For those, Java, C#, SQL, and the occasional bit of Python each are well suited for the job. There is less cognitive overhead to simply code for solving the problem, and then tweaking for minimizing performance issues and garbage-collector/resource leaks.
I have, on occasion, needed to interface with low-level hardware. In a Windows environment, C++ typically proves to be an excellent tool for the purpose. For the same purpose in Linux and DOS, I typically use "old fashioned" C if not using any APIs/libraries that would make C++ the better choice. I do not mind mind managing my heap in those cases and for C++ I can make use of RAII (stack).
This is not to say the Rust and Go (which I have not got around to trying) are bad choices for a project; however, I would not simply assume they are a silver bullet for a problem. I will also mention that sometimes a sub-optimal tool can be used for the job in a pinch. In the physical sense, I would normally use a hammer for a nail and a screwdriver for a screw-- I have used a hammer to beat in a screw and I have used the back of a screwdriver to insert a nail (and lamented the entire process).
Now, concerning ESR's position, I certainly see his reasons. Since my thoughts are difficult to articulate, I shall make a comparison to my browser: Firefox. I first started using it (under the name of Phoenix) because it was not Internet Explorer, but lighter weight than the then bloated Mozilla-branded browser. It had problems and it was lacking features, but it did provide enough customization that I was comfortable using it. Since then, missing features were added, security holes were patched, features that I had liked were removed since they were either troublesome or not popular, and it has gained bloat (likely to fulfilling features) similar to its predecessor. Other browsers are starting to look "shinier" to me. C++ is a language, but it, like application, will evolve to the niche of the target audience. Individual users, whether speaking of applications such as browsers or languages, will each have different needs and different limitations they are willing to tolerate. I prefer to focus on solving the problem, but I do not mind taking on the extra responsibilities if it brings me benefit. I have had talented co-workers that are absolutely allergic to idea of managing their own memory, but loves to "create clean elegant solutions." I have also had other talented co-workers that absolutely refuses to trust language/library provided mechanisms such as garbage collection, but they have the attention to detail and the background of knowledge to rapidly create lean and correct solutions to problems where they do have that level of control.
For the TL;DR, have a reasonable set of ability to use for the time and effort you are willing and able to invest in the niche that you can or want to fill. Find the right balance of flexibility and proficiency that matches the limited resources to invest and the opportunities available. Also, remember that ESR is a quasi-public figure with strong, sometimes bitter opinions: earnestly consider them, but take them with a grain of salt. I will also note that concerning his mention of the futility of trying to predict the future (the next big language), it is like predicting the stock market in that some things will stick around forever, but have only modest returns; however, a penny stock have the possibility of a minimal investment with a big return, but also carries a high risk of just being a useless waste of resources.
Java's speed is not the issue. Java can be translated to work as native byte code which will typically be fast enough. The real difficulty is the lack of direct memory and IO access. In the "spherical cow" case, if the hardware/firmware has all of the necessary interface code, then abstractions can just be translated by included standard libraries. This, of course, would have to rely on something like an UEFI extension that provides device services (which would be still written in a "lower-level" language). I will note that bootstrap, up until more recent, was not even done in C or C++, but in assembly and often relied on BIOS calls to get things initialized. Someone with more time on their hands than I can test this statement by writing a Java VM in Java. The other option would be providing something similar to C# "unsafe" keyword except also including direct I/O which could be applied to the encapsulating class.
The thread does beg the question, if a vacuum is easily set on fire remotely, does that mean that the vacuum really sucks? If so, although an average end-user would want a vacuum that really sucks, would they want a vacuum that sucks in this thread's context? Also, does the vacuum catching fire from indirect unapproved interfacing to it mean that the manufacturer will cast the warranty to Void? If so, does this mean that the owner will need to return to manual garbage collection? Will people even realize the gravity of the issue, and if not will they be doomed to be stretched beyond limit and then left in the dark and crushed?
Snowflakes are Liberals, no matter how much you try to change it.
Sorry but the conservatives are the real snowflakes, they get really ass-hurt when a couple of gay guys get married, they aren't permitted to impose their religious dogma on others, or a black guy gets elected to the presidency.
Snowflake status is politically and religiously agnostic. It only requires being unique like everyone else, melting when faced with real heat, and the ability to pile on top of other snowflakes to avalanche at the slightest peep. Conservatives, Liberals, Moderates, people off the artificially one dimensional political scale... Anyone can be a snowflake.
That said, the idea of nuclear reactions being triggered by lightning does not seem any more shocking (pun intended) to me than a troll trying to turn a science post into an unrelated political discussion. Remember: never directly feed a troll, instead let the trained professionals throw the troll-bait under the bridge so that it safely returns to its natural habitat...
Internet providers limits the total amount downloaded per month. In our case, it just means throttling to ISDN speeds, for some friends and acquaintances it means extra monthly fees. It is often both easier and cheaper to just buy a disc to insert into the player than to log into a streaming service, find the movie, and watch cap rapidly come near. There is also the issue of licensing. If the streaming services no longer license a movie or series, then it is gone until found again. With a disc, as long as care is taken (good storage conditions, backup, and enough technical ability to convert formats) then there is much less concern with losing access to enjoyed media due to a third party short of full disasters.
Slashdot Mirror
SCIgen referring to https://pdos.csail.mit.edu/arc... if anyone is interested
I have little doubt that Amazon had the necessary corpus to generate books including ones containing all the elements that people look for in a good book. That said, well designed and programmed robots can make reasonable food out of correct quantities of each ingredients in the correct order, but after a while it will all just begin to taste the same and be bland. In the case of the software generated stories, some of it might be accidentally good enough to make it into a literary publications if all of the vogue themes show up; however, until such time that we have "electric readers" that will read for us so that we won't have to do so, I doubt the bulk of it will make the upper ranks for the bulk of readers, at least for the first few generations of such.
Copyright and patents are still useful, but the timeframe is just purely rent-seeking. How about $10 dollars to register either for 2 years, double the previous for each 1 year extension. $10, $20, $40, $80, etc. The first few years will be cheap and a common citizen could get a foothold. Someone successful could even continue the monopoly for a few extra years. After a while, even a company such as Apple with unruly amounts of capital would give up and allow it to be public domain.
I agree for the trademarks, but would advice that it only applies to the context of use as a trademark. Trademarking a certain mouse would not prevent the character of that mouse from being used in a fictional work if not meant to represent the company.
I am not so concerned with the presence of copyright since it would be one of the few things keeping my modest work from simply being ripped off by someone with more marketing skills than I before I can even attempt to do something. I am, however, concerned with the blatant abuse. If someone is using one of my random utilities or hanging a print of one of my ugly renders on the wall 20 years from now, I would have had a chance to market it. 100 years of rent-seeking, it only encourages people to rest upon their laurels while speeding up inflation by creating undue artificial scarcity.
During a trip earlier this year to Toronto, I had visited Fort York. At least one sign referenced "Americans" plundering and burning the place. Since I doubt they would being doing it to themselves, it stands to reason they were referencing the US soldiers (history class, years ago, touched upon it as an US offensive towards the British which, while successful, wasn't quite as strategic as hoped, the museum at the site of the fort suggests the same but in less sugarcoated language, for fairly plain reasons).
Canadians and Mexicans are indeed citizens of the North American continent (or some of the citizens of the larger American continent, depending on whose map you read), but "United Statians" does not roll off the tongue as well in English, so I imagine that why it became a de facto convention. Since it is not de jure, maybe it will change, maybe it won't.
Honey-pots are not the newest idea out there; however, I might point out that actual dead-code in the binary is often removed by compilers when using higher numbers on the optimization parameter. -O1 might not remove it, but -O3 probably would (compiler dependent). I might also make note that if there is room for a buffer overflow in a library, the inside of the if statement might make an attractive jump target if is it visible to the context (probably involving setting EAX, ES, and EDI first and doing a direct far jump to the address with the current stack).
An interesting idea, but cost and risk versus benefit, I do not think it worth it.
People are using their phones increasingly more like a low-end laptop. I suspect the bargain cellphones will eventually catch up, but in the meantime, my $400 will not do the stuff that I would like to do. While I am more willing to forego doing the extra things, some people have more dollars and less restraint. We share see what the market will bear once the data is bare.
If I had mod points, I would 1-up this. I had no trouble understanding Sagan's videos when I was in grade school.
Especially since the GOTO statement is considered harmful.
Maybe that Perl of humor was the straw that broke the camel's back. That said, at the very least, Perl is still easier to read than Whitespace and easier to pronounce than INTERCAL.
I note that I feel the same "twitch" in the ears as I do just after using a firearm. I understand there is a muscle there; perhaps it is learned reflex to the visual cue.
Cause for many of us Gen-X'ers, investments haven't even broken even. Heck, we're going to be the first generation to pay more into Social Security than we'll get back.
That about sums up investing for Gen-X'ers.
I'm on the line between Gen-X'er and Millennial. After considering taxes and fees, my mutual fund near about breaks even: if sold, there would be just about enough left over to buy a grande latte and leave a decent tip. :-P
I do not think that given his potential income that "coffee shop" coffee will make or break him. On the other hand, someone that that makes far less could put that same money into savings to get a bit shy of an extra thousand a year of financial buffer, assuming that they do not choose to divert it elsewhere. Of course, the few minutes of time making his own coffee could be invested elsewhere, but if anything like my significant other, then that task is likely being "worked-in" as part of multitasking another close proximity task such as making breakfast. Before I started my career, I usually eschewed "barista made coffee" except when "treating myself" after passing my finals or getting high enough performance reviews to qualify for a bonus. Now since I make a reasonable salary, I am a bit more relaxed about "extras." That said, as per the adage, "if you want something done right, do it yourself": I like my coffee a particular way, so I still tend to brew it myself. Also, the gentleman, who probably does it purely on principle, is free to make his own choice to brew it himself just as the person that opts to pay a premium to have it done for them.
Once logged in, it takes about 10 minutes on the average Linux machine and about 15 on the average Windows 7 with Cygwin machine, including coffee breaks, to setup a private certificate authority, convert the public key cert, and copy it to a flash drive. Installing as a trusted certificate authority onto multiple Windows machines can be done rapidly using group policy, if available, or Powershell script from an admin account if not. It takes a couple of minutes per CSR to sign for the webservers and random numbers of minutes to get them into IIS. Apache is a bit faster.
That said, I have not read the details of the change, but I hope there is a readily accessible power-user option to turn it off. I worked for several companies with IT department that demands exclusive control over whatever CAs are out there but takes forever fulfilling new signing requests. I am not at one now, but through the lens of past experience, it is better to have the option to control the behaviour.
As stated, seems useful for investigation/obtaining warrant, and accuracy can be confirmed with blackbox quality assurance. On the other hand, I would refrain from using as "star evidence." That said, if I were in there shoes, I would get a my lawyers to draft up a nice pair of NDAs, get a respected university to verify the science, and get a security company to review the coding to get a pair of gold stars. It might make DA's and investigators a bit more likely to take a look.
You can scale an image to a fixed size, divide it up into parts, take a histogram of each part (Fourier transform), and shove it into an indexed database (might need to duplicate also do greyscale separately). If another image has most of the parts' histograms within a certain arbitrary percentage, the picture is probable that it is the same (no need to directly compare the original images). That, in the loosest way possible, can be considered a "hash comparison." Of course, I'm pretty sure Facebook has something more advanced than that bit of minor experimentation. Once confirmed not to be the same as a previous image, just do the normal built-in facial recognition. If unable to automatically confirm, send it to queue for an underpaid human to handle.
My "photo" is a vector art drawing of myself. Since it would have been no fun tracing a picture of myself in Inkscape, I opted for kinda-sorta manga style. I doubt a non-human can make a comparison (I can make a face humanoid and maybe even identifiable, but art isn't my forte). More than likely automatic comparison is done when it identifies photographs. For everything else, the upload likely generates a ticket for a human to pull up and click 'Yes', 'No', or 'Maybe' with 'Yes' making the system happy, 'No' suspending the account with some message, and 'Maybe' to assign the ticket to someone else or pop up a call customer-service message.
In regards to your sample:
If no reason was given, I would quickly demand:
Unless optimized inline, there would be extra work on the stack for the call, but someone unfamiliar with the idiom or the subtleties of operator precedence (or C lacking a dedicated bool type) of the original sample will have trouble. Yes, a veteran C/C++ programmer would recognize it instantly, but for crunch-time, it is not that unusual for a "loaner" to be attached to the project that normally uses a different but similar language, e.g., a skilled Java or C# person that knows what a pointer is. The latter example I provided has a clear purpose. Of course, extra help worth their salt would have the sense to look it up, assuming they aren't on their last drop of liquid motivation while burning the midnight oil.
This is related back to your mention of:
The key isn't in avoiding symbols, which you're supposed to be able to just read if you plan to do any coding at all, but to ensure formatting and sane use of whitespace makes turns a string of symbols into a set of instructions.
I am just adding that it helps to be mindful of having good naming, try to use standard libraries (except for when there is documented good reason not to do so), document when doing something esoteric (when cannot be helped), and keep the "clever" to the obfuscation contests and proof-of-concept snippets.
For the TL;DR, having good conventions (whitespace, naming, library usage, etc), inline documentations (and actual written documentation, please), reasonable communication concerning standards, and clear expectations can allow for C/C++, or most other languages for that matter, to be maintainable even in the largest of projects.
An extra after thought, the effect of a JIT compiled language may be possible at the kernel level by "faking it." While there would be limitations, upon driver installation, native image generation could be done. The kernel would need to have some sort of metadata service. Garbage collection might could be a compiler service as well, but I would opt for having the initial code compile do static analysis for reference count and just inject deallocations after the last reference. Any case that cannot be verified to remain in scope would fail the compile; the programmer would have to revert to manual allocation/deallocation possibly disguised as support libraries and use at their own risk. Additional keyword or syntactic sugar might could be used when the programmer need to delay deallocation for some reason such as performance. Use try/finally (or equivalents such as try-with-resources, using keyword) for setting allocation boundaries or accessing scarce resources. Use case would be for rapid development of drivers for cross-architecture but same kernel where the driver does not require much direct memory access or IO, but would benefit from not having a context switch.
I doubt firmware would be possible any time soon with JIT compile languages. Considering CPU speeds and memory sizes, garbage collection should be possible in firmware (after some bootstrap); however, it would be quite the waste considering the target. I would say similar for kernel drivers. User-land drivers such as certain types of human interface devices/filters, printer drivers, scanner drivers, maybe (big maybe) virtual file system drivers, or similar which would delegate its communication with hardware through an abstraction layer would at least be worth considering on systems with sufficient resources to make it worth while. Although, at that point, I would sooner call those services.
I was playing the game for a while (until schedule and loss of interest took the fun out of it). I never walked into the road without looking even once; I will admit to colliding with other players of the same game a few times in a park area when not using the augmented reality feature due to battery-life and ease of playing. On the other hand, while driving, I have had several people walk in front of me while playing (the ball toss "finger-flick" is quite distinctive) without so much as a care in the world; however, I also had similar numbers of people with both thumbs on the phone, apparently composing something, suddenly walk in front of me. My reaction time is not excellent, so if driving conditions were different, there may very have been different outcomes than me stopping and tapping my horn: we (in each case) both were lucky. While the majority were embarrassed and/or apologetic and quickly removed themselves from the right-of-way, one particular case went so far as to showing the center-finger-salute and then repeatedly hitting my vehicle while continuing to yell about it being my fault for them not successfully capturing a Bellspout, at least until I drove off. Fortunately, I never had any incident concerning a younger child running into the right-of-way while apparently playing that game. I did see, however, a few (seemingly) responsible parents taking active measures to prevent disaster, e.g., calling back child before he/she arrives at road, placing themselves between child and road, or at least intercepting the child and extracting mobile device. I strongly hope that of the parents that have children playing the game, those samples are not outliers.
Pokemon Go (and Ingress for that matter), for those immersed in the game, is indeed a distraction in the same way social media and other electronic communications are. That said, I would hesitate to ban it or provide any unnecessary movement restrictions. I am prone to be easily distracted myself; however, I can choose to forgo playing a game, texting, etc for a few moments while it is unsafe to do so. That is my responsibility. I suspect that only a tiny minority of adolescents and adults have such significant developmental, functional, or cognitive conditions that would make it impossible (or at least exceedingly difficult) to train one's self to exercise caution. For those that would have such troubles (or anyone, just for doing so), if wanting to join in playing, go with a group of friends, have fun, and be safe.
While I certainly not found of C++ myself, I can certainly code in it. As it stands, most of the problems that I have to solve on a regular basis does not involve usage of that language directly. That said, I still think there is use for it, even beyond legacy compatibility. I typically work on user-land applications, data ETL jobs, and web services. For those, Java, C#, SQL, and the occasional bit of Python each are well suited for the job. There is less cognitive overhead to simply code for solving the problem, and then tweaking for minimizing performance issues and garbage-collector/resource leaks.
I have, on occasion, needed to interface with low-level hardware. In a Windows environment, C++ typically proves to be an excellent tool for the purpose. For the same purpose in Linux and DOS, I typically use "old fashioned" C if not using any APIs/libraries that would make C++ the better choice. I do not mind mind managing my heap in those cases and for C++ I can make use of RAII (stack).
This is not to say the Rust and Go (which I have not got around to trying) are bad choices for a project; however, I would not simply assume they are a silver bullet for a problem. I will also mention that sometimes a sub-optimal tool can be used for the job in a pinch. In the physical sense, I would normally use a hammer for a nail and a screwdriver for a screw-- I have used a hammer to beat in a screw and I have used the back of a screwdriver to insert a nail (and lamented the entire process).
Now, concerning ESR's position, I certainly see his reasons. Since my thoughts are difficult to articulate, I shall make a comparison to my browser: Firefox. I first started using it (under the name of Phoenix) because it was not Internet Explorer, but lighter weight than the then bloated Mozilla-branded browser. It had problems and it was lacking features, but it did provide enough customization that I was comfortable using it. Since then, missing features were added, security holes were patched, features that I had liked were removed since they were either troublesome or not popular, and it has gained bloat (likely to fulfilling features) similar to its predecessor. Other browsers are starting to look "shinier" to me. C++ is a language, but it, like application, will evolve to the niche of the target audience. Individual users, whether speaking of applications such as browsers or languages, will each have different needs and different limitations they are willing to tolerate. I prefer to focus on solving the problem, but I do not mind taking on the extra responsibilities if it brings me benefit. I have had talented co-workers that are absolutely allergic to idea of managing their own memory, but loves to "create clean elegant solutions." I have also had other talented co-workers that absolutely refuses to trust language/library provided mechanisms such as garbage collection, but they have the attention to detail and the background of knowledge to rapidly create lean and correct solutions to problems where they do have that level of control.
For the TL;DR, have a reasonable set of ability to use for the time and effort you are willing and able to invest in the niche that you can or want to fill. Find the right balance of flexibility and proficiency that matches the limited resources to invest and the opportunities available. Also, remember that ESR is a quasi-public figure with strong, sometimes bitter opinions: earnestly consider them, but take them with a grain of salt. I will also note that concerning his mention of the futility of trying to predict the future (the next big language), it is like predicting the stock market in that some things will stick around forever, but have only modest returns; however, a penny stock have the possibility of a minimal investment with a big return, but also carries a high risk of just being a useless waste of resources.
Java's speed is not the issue. Java can be translated to work as native byte code which will typically be fast enough. The real difficulty is the lack of direct memory and IO access. In the "spherical cow" case, if the hardware/firmware has all of the necessary interface code, then abstractions can just be translated by included standard libraries. This, of course, would have to rely on something like an UEFI extension that provides device services (which would be still written in a "lower-level" language). I will note that bootstrap, up until more recent, was not even done in C or C++, but in assembly and often relied on BIOS calls to get things initialized. Someone with more time on their hands than I can test this statement by writing a Java VM in Java. The other option would be providing something similar to C# "unsafe" keyword except also including direct I/O which could be applied to the encapsulating class.
The thread does beg the question, if a vacuum is easily set on fire remotely, does that mean that the vacuum really sucks? If so, although an average end-user would want a vacuum that really sucks, would they want a vacuum that sucks in this thread's context? Also, does the vacuum catching fire from indirect unapproved interfacing to it mean that the manufacturer will cast the warranty to Void? If so, does this mean that the owner will need to return to manual garbage collection? Will people even realize the gravity of the issue, and if not will they be doomed to be stretched beyond limit and then left in the dark and crushed?
Snowflakes are Liberals, no matter how much you try to change it.
Sorry but the conservatives are the real snowflakes, they get really ass-hurt when a couple of gay guys get married, they aren't permitted to impose their religious dogma on others, or a black guy gets elected to the presidency.
Snowflake status is politically and religiously agnostic. It only requires being unique like everyone else, melting when faced with real heat, and the ability to pile on top of other snowflakes to avalanche at the slightest peep. Conservatives, Liberals, Moderates, people off the artificially one dimensional political scale... Anyone can be a snowflake.
That said, the idea of nuclear reactions being triggered by lightning does not seem any more shocking (pun intended) to me than a troll trying to turn a science post into an unrelated political discussion. Remember: never directly feed a troll, instead let the trained professionals throw the troll-bait under the bridge so that it safely returns to its natural habitat...
The HCF instruction implies "HALT" rather than "HACK," so, if executed, at least it would have stopped first.
Internet providers limits the total amount downloaded per month. In our case, it just means throttling to ISDN speeds, for some friends and acquaintances it means extra monthly fees. It is often both easier and cheaper to just buy a disc to insert into the player than to log into a streaming service, find the movie, and watch cap rapidly come near. There is also the issue of licensing. If the streaming services no longer license a movie or series, then it is gone until found again. With a disc, as long as care is taken (good storage conditions, backup, and enough technical ability to convert formats) then there is much less concern with losing access to enjoyed media due to a third party short of full disasters.