But as an OSS developer, don't you have a responsibility to fix some problem in your code? You released code so someone would use it, but the bugs make it difficult for the users to achieve their goals, which in turn may affect company profit (if said OSS is being used in a commercial setting).
If you like releasing half-baked crap that is useless to someone after they've invested real time and money, don't release such software in the first place or declare a warning stating that "support is unlikely or will take months to resolve any bugs you have or will take $$$ to fix any problem."
Bottom line, OSS projects need to specify the time it will take to resolve any issue right on their product description page. This will help corps decide whether they want to use your software or not.
I know RedHat does. If you don't have a contract with them, and you are a business, then they likely aren't going to bother with you.
But what can RedHat provide that you don't already get from the kernel developers? Suppose there's a bug in the kernel, gcc or Apache, don't these product teams already release bug fixes? Why do we need RedHat to resolve the issue?
Yes, please pass this law. Google has been making billions of other people's content. And that was okay because the content creators were hobbyists or small time blogs and these guys are usually desperate for visitors and are pushovers.
Google then used the same formula with newspapers, where unlike blogs, the copyrighted content is often right in the headline. What reader is going to visit the e-newspaper if he/she can get the content from Google News? Only a few, who're interested in the details. The rest are okay with headline and summary provided free by Google News, but which shafts advertising revenue for the newspaper which worked hard to create the story.
In this case, Google News are like a bunch of freeloading pirates.
I wonder if this law is to counter Google News displaying links to news stories and eating e-newspaper's lunch.
OTOH, since google search results is mainly links, won't it have to shut down because it can no longer display links related to your search? Without search, the internet is pretty much useless.
They should do a compromise, and allow some links to be copyrighted while others should be linkable. Private, copyrighted links should be like "http://server/privatepage.html_priv"
4 This only slightly speeds up/makes the process easier. Anything you can read can be transcribed.
The speedup is very high... any book scanned in an hour at zero cost (other than the one-time $199 scanner cost). Try transcribing manually (for example typing the contents of a book into your editor) and see how long and tedious a task that is.
If the OCR quality is as good as they say it is, the book's pdf file size will be really small (less than 50 MB).
No need to cut anything off with this scanner (if you've seen the demo youtube video). So will users just check out books from the university/public library and scan it at home? Later they can upload it to bittorrent or other sharing sites.
Is it likely this device will be banned because it allows easy circumvention of copyright laws?
Utterly wrong. Artisans (craftsmen) produce the same thing over and over again: pots, pans, bread, hand-made shoes, etc, with little innovation. Craft is just art+tech that does not change significantly.
Programming, on the other hand, involves a lot of (innovative) changes unless your job is some kind of maintenance software dev.
Software is engineering, except it is not as difficult as the traditional engineering fields like EE, mechanical etc.
The engineering isn't in writing VHDL; it's in designing the project.
Designing/architecture is separate from coding in your software/hardware language. Whether your VHDL code is synthesized to gates/flip-flops or your software compiled to CPU machine code is an overly pedantic distinction, because both pieces of technology transform/process digital data from one form into another. Before GPUs existed, 3D rendering was performed by software. Hardware just made 3D rendering faster. Anything you can do with hardware digital logic, you can also do with software.
Writing code is an art, much like mechanical drawing. It's not engineering.,
Some parts are arty, but most of it something you've learned from textbooks, other code or articles. Software is a mix of art and engineering but so is any form of engineering that brings in new technology.
I get it, if a programmer writes, a = b | c, he's not an engineer. But if a VHDL/Verilog developer writes the same a = b | c, he's suddenly an engineer, way above the lowly software code monkey, even if the code they both write does the same thing.
Both software developers and hardware Verilog/VHDL engineers write similar code, arithmetic and logic operators, conditional code, loops, variables etc. The main difference is hardware executes these statements in parallel and certain things like multiplication/division operators and arrays are not easily available, whereas software is serial in nature, and the cost per operation is way lower since code is stored as bits in flash/HDD/RAM compared to implementing transistors in an ASIC chip.
In conclusion, if software devs are not engineers then neither are digital logic engineers.
Like health inspectors checking restaurants for bad food or procedures, isn't there an independent company that can randomly test and certify USB cables?
Maybe CPU, OS and compiler tools should offer feedback to the programmer on how the program is affecting various levels of cache first, then the programmer will improve. Start with the laziness of the hardware/software vendors.
The additional commentary he adds further attacking the coder is unnecessary and hostile.
I don't agree with you that it's hostile, but let's say it is.
Also, let's say you're in charge, instead of Linus. Now replace the sentence shown below with something that is not supposedly hostile but also warns the developer and others not to repeat the same mistake.
and anybody who thinks that the above is... is just incompetent and out to lunch.
if (overflow_usub(mtu, hlen + sizeof(struct frag_hdr),... goto fail_toobig;
and anybody who thinks that the above is
(a) legible (b) efficient (even with the magical compiler support) (c) particularly safe
is just incompetent and out to lunch.
Your quote:
I frequently tell my coding minions that their code is bad and they need to rewrite it, but I don't call them names or swear at them. Instead I tell them "this is unacceptable and why."
Other than the "incompetent and out to lunch," Linus does explain the "unacceptable code and why."
There is a difference between attacking a person and critiquing their work.
He's mainly criticizing the code, but people are too thin-skinned to take good advice and instead think of it as an attack on their character. If he didn't criticize this code in a public forum, there's a good chance this same type of code will be repeated by other coders slowly weakening the code base, by making it hard for other devs to read or modify the code.
If you check out Pottering's video on youtube, he mentions something about OS X's launchd process (the init process of OS X). Launchd centrally creates sockets to remove inter-daemon dependencies that allows OS X daemons to run in parallel and therefore boot the OS quickly.
TL;DR, systemd is following the OSS tradition of copycatting proprietary technology into free software.
What's wrong with making the texts for technical courses (math, science, engineering) free?
If you can afford bus fare going to college, why can't you pay a measly $50-$70 for an important book? I agree that $180 is ridiculously high unless it's a niche/in-demand subject like law or marketing. You freeloaders are far worse than the greedy $200 textbook writers/publishers.
I have the ultimate solution, taken from the likes of OSS, lets make all of our textbooks setup with Creative Commons License that allows for competition and free digital forms for all students.
Great solution: open source. The user gets stuff for free while the creative people who produce said stuff, get nothing. This is what happens when copyright protection does not exist (authors royally shafted by consumers).
This way, the students get what they need without sacrificing (needlessly). Or the students may get a really low quality textbook because the authors are not getting paid and therefore don't care much about quality.
You need closer to 2.1, so who gets 3 and who gets 2?
Probably the upper middle class, the wealthy and the poor. The rich need to spread their wealth somehow. The poor need more children to help in their old age, chores etc.
What if the whole world adopted the 2-child policy? Won't that stabilize the population so that it won't increase or decrease in the long term? Any imbalance can be corrected with a 1-child policy (for overpopulation) or 3-child policy (for underpopulated countries).
If China has had 1-child policy since '79, why has their population increased so much? Shouldn't it have halved by now (2 parents replaced by 1 child)?
I suppose it has nothing to do with having free beta testers? Once the tech has been tested thoroughly, get ready to pay premium prices that car makers are famous for charging.
Slashdot Mirror
That's like saying, "if you don't want to be raped, don't leave your house."
But as an OSS developer, don't you have a responsibility to fix some problem in your code? You released code so someone would use it, but the bugs make it difficult for the users to achieve their goals, which in turn may affect company profit (if said OSS is being used in a commercial setting).
If you like releasing half-baked crap that is useless to someone after they've invested real time and money, don't release such software in the first place or declare a warning stating that "support is unlikely or will take months to resolve any bugs you have or will take $$$ to fix any problem."
Bottom line, OSS projects need to specify the time it will take to resolve any issue right on their product description page. This will help corps decide whether they want to use your software or not.
But what can RedHat provide that you don't already get from the kernel developers? Suppose there's a bug in the kernel, gcc or Apache, don't these product teams already release bug fixes? Why do we need RedHat to resolve the issue?
Yes, please pass this law. Google has been making billions of other people's content. And that was okay because the content creators were hobbyists or small time blogs and these guys are usually desperate for visitors and are pushovers.
Google then used the same formula with newspapers, where unlike blogs, the copyrighted content is often right in the headline. What reader is going to visit the e-newspaper if he/she can get the content from Google News? Only a few, who're interested in the details. The rest are okay with headline and summary provided free by Google News, but which shafts advertising revenue for the newspaper which worked hard to create the story.
In this case, Google News are like a bunch of freeloading pirates.
I wonder if this law is to counter Google News displaying links to news stories and eating e-newspaper's lunch.
OTOH, since google search results is mainly links, won't it have to shut down because it can no longer display links related to your search? Without search, the internet is pretty much useless.
They should do a compromise, and allow some links to be copyrighted while others should be linkable. Private, copyrighted links should be like "http://server/privatepage.html_priv"
Go to 1:13 in the video: https://www.youtube.com/watch?... . They appear to have solved page curl via the "Flattening Curve" process.
The speedup is very high... any book scanned in an hour at zero cost (other than the one-time $199 scanner cost). Try transcribing manually (for example typing the contents of a book into your editor) and see how long and tedious a task that is.
If the OCR quality is as good as they say it is, the book's pdf file size will be really small (less than 50 MB).
No need to cut anything off with this scanner (if you've seen the demo youtube video). So will users just check out books from the university/public library and scan it at home? Later they can upload it to bittorrent or other sharing sites.
Is it likely this device will be banned because it allows easy circumvention of copyright laws?
Utterly wrong. Artisans (craftsmen) produce the same thing over and over again: pots, pans, bread, hand-made shoes, etc, with little innovation. Craft is just art+tech that does not change significantly.
Programming, on the other hand, involves a lot of (innovative) changes unless your job is some kind of maintenance software dev.
Software is engineering, except it is not as difficult as the traditional engineering fields like EE, mechanical etc.
Designing/architecture is separate from coding in your software/hardware language. Whether your VHDL code is synthesized to gates/flip-flops or your software compiled to CPU machine code is an overly pedantic distinction, because both pieces of technology transform/process digital data from one form into another. Before GPUs existed, 3D rendering was performed by software. Hardware just made 3D rendering faster. Anything you can do with hardware digital logic, you can also do with software.
Some parts are arty, but most of it something you've learned from textbooks, other code or articles. Software is a mix of art and engineering but so is any form of engineering that brings in new technology.
I get it, if a programmer writes, a = b | c, he's not an engineer. But if a VHDL/Verilog developer writes the same a = b | c, he's suddenly an engineer, way above the lowly software code monkey, even if the code they both write does the same thing.
Both software developers and hardware Verilog/VHDL engineers write similar code, arithmetic and logic operators, conditional code, loops, variables etc. The main difference is hardware executes these statements in parallel and certain things like multiplication/division operators and arrays are not easily available, whereas software is serial in nature, and the cost per operation is way lower since code is stored as bits in flash/HDD/RAM compared to implementing transistors in an ASIC chip.
In conclusion, if software devs are not engineers then neither are digital logic engineers.
Would you buy a usb cable if some Joe wrote on his blog it worked for him and he rates it five stars (would buy again)?
Like health inspectors checking restaurants for bad food or procedures, isn't there an independent company that can randomly test and certify USB cables?
Maybe CPU, OS and compiler tools should offer feedback to the programmer on how the program is affecting various levels of cache first, then the programmer will improve. Start with the laziness of the hardware/software vendors.
I don't agree with you that it's hostile, but let's say it is.
Also, let's say you're in charge, instead of Linus. Now replace the sentence shown below with something that is not supposedly hostile but also warns the developer and others not to repeat the same mistake.
and anybody who thinks that the above is ... is just incompetent and out to lunch.
Your quote:
Other than the "incompetent and out to lunch," Linus does explain the "unacceptable code and why."
He's mainly criticizing the code, but people are too thin-skinned to take good advice and instead think of it as an attack on their character. If he didn't criticize this code in a public forum, there's a good chance this same type of code will be repeated by other coders slowly weakening the code base, by making it hard for other devs to read or modify the code.
Goto is good in this case since it avoids nasty 'if's in many places. You complain because you're a noob.
If you check out Pottering's video on youtube, he mentions something about OS X's launchd process (the init process of OS X). Launchd centrally creates sockets to remove inter-daemon dependencies that allows OS X daemons to run in parallel and therefore boot the OS quickly.
TL;DR, systemd is following the OSS tradition of copycatting proprietary technology into free software.
video:
https://www.youtube.com/watch?...
If you can afford bus fare going to college, why can't you pay a measly $50-$70 for an important book? I agree that $180 is ridiculously high unless it's a niche/in-demand subject like law or marketing. You freeloaders are far worse than the greedy $200 textbook writers/publishers.
Great solution: open source. The user gets stuff for free while the creative people who produce said stuff, get nothing. This is what happens when copyright protection does not exist (authors royally shafted by consumers).
Probably the upper middle class, the wealthy and the poor. The rich need to spread their wealth somehow. The poor need more children to help in their old age, chores etc.
What if the whole world adopted the 2-child policy? Won't that stabilize the population so that it won't increase or decrease in the long term? Any imbalance can be corrected with a 1-child policy (for overpopulation) or 3-child policy (for underpopulated countries).
Well, robots and automation in general, will replace need for more humans.
If China has had 1-child policy since '79, why has their population increased so much? Shouldn't it have halved by now (2 parents replaced by 1 child)?
I suppose it has nothing to do with having free beta testers? Once the tech has been tested thoroughly, get ready to pay premium prices that car makers are famous for charging.