"My hand is getting tired with all the pinching zooming. I need a good touchpad and someti a good mouse."
What!? Point me to the man who first came up with the idea of eliminating the awesome trackballs that came on ALL laptops way back when, and replaced them with the god-awful painful, slow, clumsy, and often accidentally activated touch pads, and I will find him and kill him with a rusty ax...
Whoever it was, the same mindset spilled over into the keyboard as well... Honestly, I can find hundreds of awesome super-compact keyboards, smaller than what's found on most laptops/netbooks, yet laptop keyboards are always horrid for some reason. No, dammit, moving the backslash key all over the place, and making severl of the keys smaller is NOT OK.
In short, I'd call a touchscreen a major upgrade... Though it wouldn't be hard for a halfway decent input device to be a huge upgrade as well. Hell, don't like trackballs? Why not a simple joystick? Cursor keys ala X11? ANYTHING to get rid of those %&#*$"# touchpads...
Your shell can do math just fine, no, you really shouldn't be forking for it. And you really shouldn't be putting leading zeros in your variables: your data should always be pure and normal. You can format it appropriately when you use it (eg. with printf).
Thank you for the universal truths. You certainly wouldn't need to back that up with reasoning or anything. You should be doing things the much more complex way, because it is POSSIBLE to do so. Wonderful reasoning.
Your own arguments defeat ksh for the same reason they defeat bash
Quite the opposite. Stick with ksh88 syntax, and avoid just a handful that bash doesn't understand/agree with, and you can pretty well guarantee your script will work damn near anywhere.
Of course, you've again jumped to the answer without any reasoning, so it must be true...
On the matter of "standard equivalent that works as well", when your hashbang is/bin/bash, you'd be silly to break your script's consistency and go for a "portable" solution (whose portability is broken by your hashbang) whose bash-equivalent has nothing but advantages.
You'd be a bad script writer for using #!/bin/bash. Most OSes (outside of Linux) won't have bash under/bin/. All the *BSDs will be/usr/local/bin/bash. Sadly, I haven't seen a consistent location for env (/bin or/usr/bin ?) either, so you're at best 50/50 there.
My point in short, changing the hash bang in a script is trivial. Replacing all bash'isms is NOT. "Porting" to a system with only ksh, perhaps where bash won't compile, can be trivial, or a nightmare, all for the sake of a few features that are just as easy to use in a more compatible way, rather than the bash way...
sometimes doing things correctly and supporting all the edge cases is indeed more complex than supporting only filenames that contain no spaces or glob characters.
No. File name handling is a major issue. The trivial issues I'm talking about are more along the lines of inserting a single line in several files... Trivial, right? Not if you follow the recommendation of piping printf to ed, nested inside a find, awkwardly escaped, incidentally with poor performance (one file at a time). This twisted usage suggested because (whoever) likes that it will transparently handle symlinks to files...
Or perhaps see the recommendation to "never, ever use seq". Even though "sed -w" will automatically give you the right number of leading zeros you need, every time, and the bash'ism equivalent requires KNOWING how many leading zeros you will need, and/or writing a complex function to replace this single simple usage. Why? See above...
Frankly, if you think you need bash'isms to write reliable shell scripts, you need to expand your horizons greatly, because it's really more the opposite. You're far more likely to run into bash'ism behavior changing, than to find a case where you NEED bash'isms to write a resilient script.
There are many existing countries where you'd be happier. Why not go there instead of trying to make this one a clone of those? I want there to be at least one remaining nation with a minimal government, such that anyone who doesn't like that can go to any other nation on earth. Is that so much to ask?
Asking half the population of a continent to leave... because they're out-voting your side, and stopping you from remaking this country into your own economic fantasy land... no, that's not unreasonable at all! In fact that's much more reasonable than YOU getting the hell out, and moving to one of many other countries with an extremely weak government that doesn't even bother trying to pass any social laws or protections.
The thing that strikes me about that wiki is the absolute obsession. bash' isms are pushed, HARD, on the unsuspecting, even when there is damn near always a standard equivalent that works well. Also, the constant hammer of recommending the most complex of all possible options, the one that proves impossibly complex when stuck in a pipeline, all because it may behave nominally better in a given border case on a single release of a single less popular OS. It's extremely pedantic, and even then, in ways no two pendants would agree on...
It's sad that "bash" shell scripting guides are so prevalent. Doubly so when it's even bash 4' isms! Even if you wish to have nice features at your finger tips, ksh93 has been gpl for some time, and oksh/mksh are quite superior to bash in several ways (line wrapping in bash has never worked right) and being far tinier and more portable is quite nice. Preferences not withstanding, you close off a big world to yourself, which you will regret later, when you depend on proprietary features in your coding/scripting in general.
I think the difference here is that I would NEVER put an SSH key on my phone in the first place. In fact you wouldn't be able to either, since you need to VPN in with 2-factor auth to even access any of my servers.
Now, back to the case of what we're actually talking about (FTP), someone having access to your non-privileged account, and your files is a minor issue. Most people don't have anything that's really all that important. We're really greatly improving security, since passwords are no longer transferred in plain text. In fact needing access to the key, and not being able to brute force a password is a huge improvement as well. Try to keep some perspective.
but that requires generating keypairs and copying the public keys around. If you're nitpicky about having separate keypairs on each SSH client machines (and you really should be!), and you have 20 hosts, then you'd have to copy 19 public keys to each machine.
This is nonsensical.
A) You should have ONE ssh key, which is password-protected. B) You start ssh-agent, ssh-add your key, and use agent forwarding (enabled by default). You can now jump around between any and all SSH/SFTP servers freely.
You only feel the need to have 19 ssh keys if you don't know about ssh-agent (putty-agent on Windows). Meanwhile, my single public SSH key is on THOUSANDS of extremely important servers...
People have said the same thing, many times before. Coal gasification was supposed to be price-competitive when gasoline reached $2.00/gal, but here we are. I have no doubt we'll see refining oil from atmospheric carbon + water also go nowhere long after we've exceeded the magic $8 price point.
most passive ftp clients use the standard sockets call for system-assigned source port like any other IP application.
What system calls are made is completely irrelevant. What matters is that,
For http/SFTP, I can just add an ACL to allow one tcp port though. Client side, server side, whatever.
For active mode FTP, I can similarly allow tcp port 21, and also tcp port 20 as the source for return connections. Annoying, but workable.
But for passive mode FTP, I have to allow the client to make connections to any and all ephemeral ports. Yes, when you're running the FTP server you can USUALLY restrict it to a modest range of a few thousand TCP ports.
If passive mode FTP had simply standardized on only one extra listening port, and/or behaved like every other TCP allication out there, all would be good with the world. Instead, we have multiple modes, none of which always work...
while the average (that is not geek) windows/osx user is going to have trouble with scp, most ftp clients are easy enough to comprehend for them...
What? SFTP is simpler to use than FTP. And don't forget all the GUI clients like WinSCP or FIleZilla which do SFTP, SCP, and FTP all in the same model. There's no added complexity at all, you get the same username / password prompt as you do with FTP.
Meanwhile with TLS FTP, you have to deal with all the same FTP active/passive, numerous ports, and possibly other negotiation issues. It isn't simple, that's why NOBODY uses it.
You've completely missed the context. We're talking about ftp here... INTERACTIVE command-line clients, which can browse the heirachy, download and upload files, launch a file editor, prompt for authentication, etc.
Hell, wget can't upload any files at all, while wput is hard to find anywhere, and buggy as hell to boot. Doesn't do jack with http anyhow.
Meanwhile, every system out there has a simple "ftp" program installed which trivially does all of the above.
What advantage is there to having unencrypted file transfers?
1) Less than 1/2 the bandwidth usage. 2) Dramatically lower server processing requirements.
I don't think I've ever owned a PC that couldn't encrypt/decrypt at the speed of my WAN connection, and my fastest LAN transfer uses less than 15% of my CPU.
That's just fine and well for your PC, but the other end is much uglier... a major FTP/HTTP server is going to crawl under the load of encrypting data for hundreds and thousands of PCs.
I agree with you, right up until you claim perl is remotely close to bourne syntax...
Sure, some of the latest bash' isms look perl' ish, but that's about it. Perl gets -1 trillion points for the very existence of the "unless" function... You could read through an entire program, line by line, looking for any possible issues, only to find an unless statement that is always false... Meaning none of that is ever executed anyhow.
Perl is notorious for how horrendously unreadable and un maintainable it is. Bourne has nothing of the sort to contend with.
Nothing's perfect, but if it's not massive, and not performance critical, I'm writing it in bourne.
Since nuclear power is essentially set back at least a decade, anything that gets us free from coal and oil is a must have, not just for global warming, but to prevent countries having to go to war for their dino juice stakes.
Why do so many people use this nonsense arguement?
Oil != Coal
Getting rid of one won't rid us of the other. It won't even help just a little. We could eliminate all coal mining in favor of solar, yet we'd still go to war over oil supplies. The price of electricity could fall drastically, and $5/gallon oil would still be in high demand.
If you want to get off oil, do it, right now. Waiting until we get off coal is unnecessary, and merely an excuse to do nothing.
Active mode FTP is hideous where NAT is involved, because it requires the server to initiate an active connection to the client.
Sadly, passive mode is horrible because it uses ephemeral ports on both ends, so you have no way to easily allow ftp and nothing else.
This leaves you in the situation of absolutely requiring an ftp proxy, because you only allow active mode on site, but passive mode is needed to get off site...
FTP is a nightmare. It has only remained because A) no command line HTTP file transfer clients ever sprang up, and B) The OpenSSH folks didn't allow you to choose unencrypted data connections for "anonymous" and non sensitive data. Either of the two would blow FTP out of the water so fast it would make your head spin. FTP is just that horrible.
I will say, however, it's very simple. It only becomes complex when you try to tack on all the failed protocols that attempt to add encryption to it, poorly, and still have the original bad design of FTP. SFTP is worlds better by comparison.
Your memory is suspect... NT 4.0 had Alpha support at launch. It was even in Windows 2000 RCs, and only dropped come release time, though I do recall Compaq selling Alpha servers with 2000 despite that.
Area 51 is chock full of advanced but terrestria technology.
No. Paste tense only, please. Area 51 is a chemical dump. An industrial hazard. It has been largely shut down for many years now.
But unless there's some lovely scifi physics waiting out there for us, space travel seems like it'll be awfully damned expensive and complicated.
Before the invention of the steam engine, traveling a thousand miles would have seemed dammed near impossible too. A working Boussard ramjet or other sustainable fusion reactor would completely redefine space travel, and make intergallactic travel quite possible.
NOTE: I didn't say you'd hop on a ship and head to the far end of the universe, as space westerns depict future space travel. Instead it's likely to be something like solar-system hopping space colonization... a few hundred people get on a ship, fly for 5 years to reach the next solar system, and build up society and industry just in time for a number of their kids to grow up and get on the next ship going to the next solar system.
I think the odds of alien life in this universe are very good; I think the odds for intelligent life are also good.
I don't. I believe it's far too covenient that with a massive number of unknowns, and no evidence, that people (scientists included) are reaching the conclusion that there MUST be intelligent life out there, somewhere. It strikes me as purely wishful thinking that then gets justified despite no evidence objectively supporting it. It seems everyone just WANTS the Deus Ex Machina to be out there, ready to come along and solving all our problems, and giving us an easy shortcut to everything we hope to accomplish.
They assumed he must have died right away. Just as everyone else is assuming would happen. Yet he was alive and in and out of consciousness for quite some time.
Humans need some animals. Bees for example. We aren't going to let them go extinct just because some natural selection pressure may be at fault, rather than human behavior.
And for the animals humans don't benefit from... Why SHOULD we save them? And no the answer can't be recursive logic or hypothetical ignorance of the consequences.
In short, whether humans are to blame or not is of no consequence, in any way, to anyone or anything.
This is nonsense. Refineries have a tremendous amount of control. It's true there would always be some gasoline in the mix, but it could be a tiny percentage, if desired. I don't see them giving away butanol, propane, etc at fire sale prices...
Nobody rains down money on education. Democrats just substantially reduce the bleeding.
Adjust public school funding today for inflation and population growth, with that of 40 years ago, and you'll see a desert, devoid of rain, with non stop bleeding.
It can be hard to answer if humans are somehow the underlying cause. And what happens when we find that bees are just naturally going extinct? Natural or no, if we loose them, we're royally screwed, and can't wait millions of years for the next pollinators to evolve.
As always, the planet will be just fine, it's just people that will be going away...
Right because science is all about being paralyzed with endless uncertainty, and making no conclusions until you're sure your statement is utterly infallible for all time.
Slashdot Mirror
What!? Point me to the man who first came up with the idea of eliminating the awesome trackballs that came on ALL laptops way back when, and replaced them with the god-awful painful, slow, clumsy, and often accidentally activated touch pads, and I will find him and kill him with a rusty ax...
Whoever it was, the same mindset spilled over into the keyboard as well... Honestly, I can find hundreds of awesome super-compact keyboards, smaller than what's found on most laptops/netbooks, yet laptop keyboards are always horrid for some reason. No, dammit, moving the backslash key all over the place, and making severl of the keys smaller is NOT OK.
In short, I'd call a touchscreen a major upgrade... Though it wouldn't be hard for a halfway decent input device to be a huge upgrade as well. Hell, don't like trackballs? Why not a simple joystick? Cursor keys ala X11? ANYTHING to get rid of those %&#*$"# touchpads...
Thank you for the universal truths. You certainly wouldn't need to back that up with reasoning or anything. You should be doing things the much more complex way, because it is POSSIBLE to do so. Wonderful reasoning.
Quite the opposite. Stick with ksh88 syntax, and avoid just a handful that bash doesn't understand/agree with, and you can pretty well guarantee your script will work damn near anywhere.
Of course, you've again jumped to the answer without any reasoning, so it must be true...
Thanks for the straw man! Nice.
You'd be a bad script writer for using #!/bin/bash. Most OSes (outside of Linux) won't have bash under /bin/. All the *BSDs will be /usr/local/bin/bash. Sadly, I haven't seen a consistent location for env (/bin or /usr/bin ?) either, so you're at best 50/50 there.
My point in short, changing the hash bang in a script is trivial. Replacing all bash'isms is NOT. "Porting" to a system with only ksh, perhaps where bash won't compile, can be trivial, or a nightmare, all for the sake of a few features that are just as easy to use in a more compatible way, rather than the bash way...
No. File name handling is a major issue. The trivial issues I'm talking about are more along the lines of inserting a single line in several files... Trivial, right? Not if you follow the recommendation of piping printf to ed, nested inside a find, awkwardly escaped, incidentally with poor performance (one file at a time). This twisted usage suggested because (whoever) likes that it will transparently handle symlinks to files...
Or perhaps see the recommendation to "never, ever use seq". Even though "sed -w" will automatically give you the right number of leading zeros you need, every time, and the bash'ism equivalent requires KNOWING how many leading zeros you will need, and/or writing a complex function to replace this single simple usage. Why? See above...
Frankly, if you think you need bash'isms to write reliable shell scripts, you need to expand your horizons greatly, because it's really more the opposite. You're far more likely to run into bash'ism behavior changing, than to find a case where you NEED bash'isms to write a resilient script.
Asking half the population of a continent to leave... because they're out-voting your side, and stopping you from remaking this country into your own economic fantasy land... no, that's not unreasonable at all! In fact that's much more reasonable than YOU getting the hell out, and moving to one of many other countries with an extremely weak government that doesn't even bother trying to pass any social laws or protections.
The thing that strikes me about that wiki is the absolute obsession. bash' isms are pushed, HARD, on the unsuspecting, even when there is damn near always a standard equivalent that works well. Also, the constant hammer of recommending the most complex of all possible options, the one that proves impossibly complex when stuck in a pipeline, all because it may behave nominally better in a given border case on a single release of a single less popular OS. It's extremely pedantic, and even then, in ways no two pendants would agree on...
It's sad that "bash" shell scripting guides are so prevalent. Doubly so when it's even bash 4' isms! Even if you wish to have nice features at your finger tips, ksh93 has been gpl for some time, and oksh/mksh are quite superior to bash in several ways (line wrapping in bash has never worked right) and being far tinier and more portable is quite nice. Preferences not withstanding, you close off a big world to yourself, which you will regret later, when you depend on proprietary features in your coding/scripting in general.
I think the difference here is that I would NEVER put an SSH key on my phone in the first place. In fact you wouldn't be able to either, since you need to VPN in with 2-factor auth to even access any of my servers.
Now, back to the case of what we're actually talking about (FTP), someone having access to your non-privileged account, and your files is a minor issue. Most people don't have anything that's really all that important. We're really greatly improving security, since passwords are no longer transferred in plain text. In fact needing access to the key, and not being able to brute force a password is a huge improvement as well. Try to keep some perspective.
Data just isn't nearly as compressible after encryption.
This is nonsensical.
A) You should have ONE ssh key, which is password-protected.
B) You start ssh-agent, ssh-add your key, and use agent forwarding (enabled by default). You can now jump around between any and all SSH/SFTP servers freely.
You only feel the need to have 19 ssh keys if you don't know about ssh-agent (putty-agent on Windows). Meanwhile, my single public SSH key is on THOUSANDS of extremely important servers...
People have said the same thing, many times before. Coal gasification was supposed to be price-competitive when gasoline reached $2.00/gal, but here we are. I have no doubt we'll see refining oil from atmospheric carbon + water also go nowhere long after we've exceeded the magic $8 price point.
What system calls are made is completely irrelevant. What matters is that,
For http/SFTP, I can just add an ACL to allow one tcp port though. Client side, server side, whatever.
For active mode FTP, I can similarly allow tcp port 21, and also tcp port 20 as the source for return connections. Annoying, but workable.
But for passive mode FTP, I have to allow the client to make connections to any and all ephemeral ports. Yes, when you're running the FTP server you can USUALLY restrict it to a modest range of a few thousand TCP ports.
If passive mode FTP had simply standardized on only one extra listening port, and/or behaved like every other TCP allication out there, all would be good with the world. Instead, we have multiple modes, none of which always work...
What? SFTP is simpler to use than FTP. And don't forget all the GUI clients like WinSCP or FIleZilla which do SFTP, SCP, and FTP all in the same model. There's no added complexity at all, you get the same username / password prompt as you do with FTP.
Meanwhile with TLS FTP, you have to deal with all the same FTP active/passive, numerous ports, and possibly other negotiation issues. It isn't simple, that's why NOBODY uses it.
You've completely missed the context. We're talking about ftp here... INTERACTIVE command-line clients, which can browse the heirachy, download and upload files, launch a file editor, prompt for authentication, etc.
Hell, wget can't upload any files at all, while wput is hard to find anywhere, and buggy as hell to boot. Doesn't do jack with http anyhow.
Meanwhile, every system out there has a simple "ftp" program installed which trivially does all of the above.
1) Less than 1/2 the bandwidth usage.
2) Dramatically lower server processing requirements.
That's just fine and well for your PC, but the other end is much uglier... a major FTP/HTTP server is going to crawl under the load of encrypting data for hundreds and thousands of PCs.
I agree with you, right up until you claim perl is remotely close to bourne syntax...
Sure, some of the latest bash' isms look perl' ish, but that's about it. Perl gets -1 trillion points for the very existence of the "unless" function... You could read through an entire program, line by line, looking for any possible issues, only to find an unless statement that is always false... Meaning none of that is ever executed anyhow.
Perl is notorious for how horrendously unreadable and un maintainable it is. Bourne has nothing of the sort to contend with.
Nothing's perfect, but if it's not massive, and not performance critical, I'm writing it in bourne.
Try again... I'm not taking a flight to the data center to plug in a usb flash drive.
Why do so many people use this nonsense arguement?
Oil != Coal
Getting rid of one won't rid us of the other. It won't even help just a little. We could eliminate all coal mining in favor of solar, yet we'd still go to war over oil supplies. The price of electricity could fall drastically, and $5/gallon oil would still be in high demand.
If you want to get off oil, do it, right now. Waiting until we get off coal is unnecessary, and merely an excuse to do nothing.
Active mode FTP is hideous where NAT is involved, because it requires the server to initiate an active connection to the client.
Sadly, passive mode is horrible because it uses ephemeral ports on both ends, so you have no way to easily allow ftp and nothing else.
This leaves you in the situation of absolutely requiring an ftp proxy, because you only allow active mode on site, but passive mode is needed to get off site...
FTP is a nightmare. It has only remained because A) no command line HTTP file transfer clients ever sprang up, and B) The OpenSSH folks didn't allow you to choose unencrypted data connections for "anonymous" and non sensitive data. Either of the two would blow FTP out of the water so fast it would make your head spin. FTP is just that horrible.
I will say, however, it's very simple. It only becomes complex when you try to tack on all the failed protocols that attempt to add encryption to it, poorly, and still have the original bad design of FTP. SFTP is worlds better by comparison.
Your memory is suspect... NT 4.0 had Alpha support at launch. It was even in Windows 2000 RCs, and only dropped come release time, though I do recall Compaq selling Alpha servers with 2000 despite that.
No. Paste tense only, please. Area 51 is a chemical dump. An industrial hazard. It has been largely shut down for many years now.
Before the invention of the steam engine, traveling a thousand miles would have seemed dammed near impossible too. A working Boussard ramjet or other sustainable fusion reactor would completely redefine space travel, and make intergallactic travel quite possible.
NOTE: I didn't say you'd hop on a ship and head to the far end of the universe, as space westerns depict future space travel. Instead it's likely to be something like solar-system hopping space colonization... a few hundred people get on a ship, fly for 5 years to reach the next solar system, and build up society and industry just in time for a number of their kids to grow up and get on the next ship going to the next solar system.
I don't. I believe it's far too covenient that with a massive number of unknowns, and no evidence, that people (scientists included) are reaching the conclusion that there MUST be intelligent life out there, somewhere. It strikes me as purely wishful thinking that then gets justified despite no evidence objectively supporting it. It seems everyone just WANTS the Deus Ex Machina to be out there, ready to come along and solving all our problems, and giving us an easy shortcut to everything we hope to accomplish.
They assumed he must have died right away. Just as everyone else is assuming would happen. Yet he was alive and in and out of consciousness for quite some time.
Humans need some animals. Bees for example. We aren't going to let them go extinct just because some natural selection pressure may be at fault, rather than human behavior.
And for the animals humans don't benefit from... Why SHOULD we save them? And no the answer can't be recursive logic or hypothetical ignorance of the consequences.
In short, whether humans are to blame or not is of no consequence, in any way, to anyone or anything.
This is nonsense. Refineries have a tremendous amount of control. It's true there would always be some gasoline in the mix, but it could be a tiny percentage, if desired. I don't see them giving away butanol, propane, etc at fire sale prices...
Not on/. We won't. Don't forget your audience. It's nearly all linux extremists.
Nobody rains down money on education. Democrats just substantially reduce the bleeding.
Adjust public school funding today for inflation and population growth, with that of 40 years ago, and you'll see a desert, devoid of rain, with non stop bleeding.
It can be hard to answer if humans are somehow the underlying cause. And what happens when we find that bees are just naturally going extinct? Natural or no, if we loose them, we're royally screwed, and can't wait millions of years for the next pollinators to evolve.
As always, the planet will be just fine, it's just people that will be going away...
Right because science is all about being paralyzed with endless uncertainty, and making no conclusions until you're sure your statement is utterly infallible for all time.