Anyone building high voltage connectors builds them with a sheath around the conductors. No one's going to be building connectors that look like household appliance plugs.
There are any number of ways to design a plug such that it's not physically possible to touch the hot wires before the ground is solidly engaged. Adding an electronic handshake to confirm would make it even safer.
In most respects, electrical safety is far easier to enforce then something like petrol pump safety.
But I could pretty easily (if I wanted to), go to a gas station, accidentally pump gas all over my pants and walk away uninjured,
This is not actually true.
Gasoline is only really "safe" to spill on yourself because it generally evaporates quickly - far faster then it can be absorbed by the skin (hence the "cool" feeling if you do happen to do this). This only holds true if the quantity remains small though, and it's not absorbed into something - like your pants.
It would be a supremely bad idea to continue wearing pants that have absorbed a significant quantity of gasoline and then just waiting for them to dry out. Yes you're "uninjured", but you're going to risk seriously poisoning yourself from the skin absorption (something like a chemical burn) and were you to then get in an enclosed space you could be overcome by fumes.
Talking about "spilling" a little electricity falls into the same category of risk as gasoline - you pretty much can't without concocting a scenario where someone does something about as stupid. A properly designed conductive charge connector isn't going to enable you to manually make contact without going to some effort to do so, and can be setup to prevent charging until a positive lock with the vehicle is established. The short circuit condition can be killed very quickly as well.
Also because ISPs in most parts of the world have made legitimate P2P impractical (the continued existence of asymmetric plans and download caps etc.).
Also a ton of those examples were talking about environments which are much harsher then LEO.
Most "space ready" components are actually more "launch ready". I mean yes space has certain dangers and conditions, but it's a pretty predictable, stable environment. Rocket launches are the hard part where subtle failures cause big explosions.
Whether you add another 8 bits or another 256 bits, the reality of the problem is the fact that for anything to work you have to upgrade pretty much the entire routing infrastructure of the internet, and solve a chicken and egg problem.
That the addresses are complicated is irrelevant. There's nothing particularly nice about an IP address, we're just used to them.
No, but if you consider a menu bar, it very easily let's you access the whole thing from the keyboard. It even tells you in the UI *how* to do this. Not so for the Ribbon.
You could argue it's very much the fault of the prevalence of NAT too - most of the time it simply wasn't a problem, but not for any good underlying reason (i.e. USB key or drive-by download type attacks then propagating via RPC vulnerabilities on internal networks).
I mean, silver-lining of NAT, whatever, but I don't think it's been remotely worth it.
It's hard to beat the simplicity modern Word offers for dropping in graphics and graphs. Being able to copy+paste or just drag things in is a hell of a lot easier then saving separate files for things.
I can hit the menu bar at the top of the screen 100% of the time without even looking. Move the mouse upwards and it stops automatically when you get to the menu bar. It's a far easier target to hit than a thin menu attached to each window.
Conversely when - ever - do you need to click on a menu without looking at the screen to do it?
Why not simply include a high-res surface mapping element to the mission? Any probe needs to insert into Europa orbit first, why not design the mission to allow the landing sight to be selected after a detailed surface map is constructed?
Also worth considering is that von Neumann type devices are essentially "life" anyway. It seems quite likely to me that it may not be possible to destroy the galaxy with them, because after a few hundred generations they themselves end up evolving into something more benign (not to mention, interstellar distance itself would be a selective pressure in such a case to not trashing every planet they land on) - or burn themselves out with some unfortunate local resource deficiencies.
It also discounts efficiency improvements. We've been very good at progressively doing more with less. Our energy use will very likely plateau, simply because technology requires it (see modern CPUs).
It also pre-supposes that anyone with the resources to do that sort of thing would bother to do it.
Western civilization has declining birth rates which means our population will stabilize - and space travel is a lousy way to beat over-population anyway (assuming of course, it doesn't turn out you can just build stargates or whatever). But even then, chances are we wouldn't need to expand to more then a few planets to have enough room for everyone since if everyone comes up to Western-level consumption, then 9-12 billion people would be all the humans we'd ever need to provide for.
So considering that, there's a decent chance that, even if another civilization had a very well funded space program and tons of resources at their disposal, the volume of actual exploration might be quite low, and their interest in setting up additional colonies might be very limited (I'm sure some would want to do it, but when you've successfully made your homeworld a paradise, "why leave" is a trickier question).
Of course, maybe I'm wrong, someone did do it, and it just turns the von Neumann probes take longer then anticipated.
No reason for it to be so dark though. It's entirely possible civilization, after a few hundred million years, ends up heading off to somewhere else to do interesting things. Or in a world without FTL, slows itself down so it can actually explore the universe in a (subjective) reasonable amount of time that makes it very much unreasonable for us to observe.
He's referring to the chance of intelligent life similar to humans coming into existence at least once.
The more interesting question is what is the follow probability of it happening more then once after that - which is what you're referring to (and I very much agree with).
You've put an improved "first time user experience" in opposition to "stability, performance, security" a lot of times where it doesn't need to be.
For example:
And then someone will replace regex with a wizard ("How many letters? What kind of letters? Should they be preceded by a space or followed by a tab? Do you want colons and double quotes with that?").
These are not mutually exclusive options. In fact, we would benefit a great deal from exposing to the user a tool which guided them through regex construction, and then showed them the regex as it was built up (and included say, tool-tips which would explain each part - I have such an app which I use for just this purpose).
I feel the "mutual exclusivity" is very much the crux of the matter though - the DE's we're all complaining about, lately, seem to be very much focused on putting these ideas in opposition.
The "simplified" interface should make way to the normal or advanced interface with an absolute minimum of effort - but I see nothing inherently wrong with trying to design the UI so that it slowly educates users about more advanced concepts in Linux/Unix (especially because, it's self-advertising as well - I've amazed people when I see them doing some repetitive task, and show them how a relatively simple shell script can automate it - but what would be even better, would be if we had a UI which got them into that mindset in the first place).
I'm always looking for a "best" way of doing something, but a lot of those choices would essentially be my "simplest" fallbacks when I want stuff to just work.
LXDE, interestingly, also has a much more "Unix" design philosophy behind it - self-contained apps which do a single-task very very well.
My first exposure to it was entirely without realizing when I was trying to figure out if I could put a touchscreen interface together using just "panel" apps that ran the appropriate commands for different tasks. The LX music player delivered.
IMO it represents a much more correct design philosophy for building a useful Linux desktop under open-source development, especially when you note that a lot of the current desktops that have pretty much gone to hell, went there around the same time they decided they were going to "integrate" everything.
The "four corners" idea gets implemented to the detriment of other ideas though. It's just not that important, and its probably the number 1 thing I despise about Mac OS (and by extension, Unity) - constantly swapping around my application menus at the top of the screen.
"Mean time to console" should probably be a UI metric - how long after installing a new DE does the average user find they need to open up a text configuration file or run a console command to fix or customize something.
Most of the forced paradigm shifts in recent years are notable because there's absolutely nothing about them which requires a forced paradigm shift. The Office "Ribbon" comes to mind (the entire UI framework easily supports traditional style toolbars, there's no reason a "2003 mode" couldn't have been left in).
Slashdot Mirror
Anyone building high voltage connectors builds them with a sheath around the conductors. No one's going to be building connectors that look like household appliance plugs.
There are any number of ways to design a plug such that it's not physically possible to touch the hot wires before the ground is solidly engaged. Adding an electronic handshake to confirm would make it even safer.
In most respects, electrical safety is far easier to enforce then something like petrol pump safety.
But I could pretty easily (if I wanted to), go to a gas station, accidentally pump gas all over my pants and walk away uninjured,
This is not actually true.
Gasoline is only really "safe" to spill on yourself because it generally evaporates quickly - far faster then it can be absorbed by the skin (hence the "cool" feeling if you do happen to do this). This only holds true if the quantity remains small though, and it's not absorbed into something - like your pants.
It would be a supremely bad idea to continue wearing pants that have absorbed a significant quantity of gasoline and then just waiting for them to dry out. Yes you're "uninjured", but you're going to risk seriously poisoning yourself from the skin absorption (something like a chemical burn) and were you to then get in an enclosed space you could be overcome by fumes.
Talking about "spilling" a little electricity falls into the same category of risk as gasoline - you pretty much can't without concocting a scenario where someone does something about as stupid. A properly designed conductive charge connector isn't going to enable you to manually make contact without going to some effort to do so, and can be setup to prevent charging until a positive lock with the vehicle is established. The short circuit condition can be killed very quickly as well.
Doesn't Twitter use BitTorrent to push code updates out amongst it's data centers?
(One of the greatest uses of BitTorrent I've heard of actually).
Also because ISPs in most parts of the world have made legitimate P2P impractical (the continued existence of asymmetric plans and download caps etc.).
If you suddenly start using encryption, it'll throw up a big flag.
If you've always been using strong encryption, then there's no change in behavior to be noticed.
Using no encryption is a bad idea, since data mining digital content is straightforward and "talking in code" is going to be picked up.
OTR messaging is great, but it's not practical to use with email (the notes on its development discuss this).
It is a spectacularly thorough system.
You do realize that space's principle property is the lack of any atmosphere right?
Also a ton of those examples were talking about environments which are much harsher then LEO.
Most "space ready" components are actually more "launch ready". I mean yes space has certain dangers and conditions, but it's a pretty predictable, stable environment. Rocket launches are the hard part where subtle failures cause big explosions.
This has precisely nothing to do with it.
Whether you add another 8 bits or another 256 bits, the reality of the problem is the fact that for anything to work you have to upgrade pretty much the entire routing infrastructure of the internet, and solve a chicken and egg problem.
That the addresses are complicated is irrelevant. There's nothing particularly nice about an IP address, we're just used to them.
No, but if you consider a menu bar, it very easily let's you access the whole thing from the keyboard. It even tells you in the UI *how* to do this. Not so for the Ribbon.
You could argue it's very much the fault of the prevalence of NAT too - most of the time it simply wasn't a problem, but not for any good underlying reason (i.e. USB key or drive-by download type attacks then propagating via RPC vulnerabilities on internal networks).
I mean, silver-lining of NAT, whatever, but I don't think it's been remotely worth it.
It's hard to beat the simplicity modern Word offers for dropping in graphics and graphs. Being able to copy+paste or just drag things in is a hell of a lot easier then saving separate files for things.
And yeah, most journals offer a Word template.
I can hit the menu bar at the top of the screen 100% of the time without even looking. Move the mouse upwards and it stops automatically when you get to the menu bar. It's a far easier target to hit than a thin menu attached to each window.
Conversely when - ever - do you need to click on a menu without looking at the screen to do it?
Why not simply include a high-res surface mapping element to the mission? Any probe needs to insert into Europa orbit first, why not design the mission to allow the landing sight to be selected after a detailed surface map is constructed?
Also worth considering is that von Neumann type devices are essentially "life" anyway. It seems quite likely to me that it may not be possible to destroy the galaxy with them, because after a few hundred generations they themselves end up evolving into something more benign (not to mention, interstellar distance itself would be a selective pressure in such a case to not trashing every planet they land on) - or burn themselves out with some unfortunate local resource deficiencies.
It also discounts efficiency improvements. We've been very good at progressively doing more with less. Our energy use will very likely plateau, simply because technology requires it (see modern CPUs).
It also pre-supposes that anyone with the resources to do that sort of thing would bother to do it.
Western civilization has declining birth rates which means our population will stabilize - and space travel is a lousy way to beat over-population anyway (assuming of course, it doesn't turn out you can just build stargates or whatever). But even then, chances are we wouldn't need to expand to more then a few planets to have enough room for everyone since if everyone comes up to Western-level consumption, then 9-12 billion people would be all the humans we'd ever need to provide for.
So considering that, there's a decent chance that, even if another civilization had a very well funded space program and tons of resources at their disposal, the volume of actual exploration might be quite low, and their interest in setting up additional colonies might be very limited (I'm sure some would want to do it, but when you've successfully made your homeworld a paradise, "why leave" is a trickier question).
Of course, maybe I'm wrong, someone did do it, and it just turns the von Neumann probes take longer then anticipated.
No reason for it to be so dark though. It's entirely possible civilization, after a few hundred million years, ends up heading off to somewhere else to do interesting things. Or in a world without FTL, slows itself down so it can actually explore the universe in a (subjective) reasonable amount of time that makes it very much unreasonable for us to observe.
He's referring to the chance of intelligent life similar to humans coming into existence at least once.
The more interesting question is what is the follow probability of it happening more then once after that - which is what you're referring to (and I very much agree with).
You've put an improved "first time user experience" in opposition to "stability, performance, security" a lot of times where it doesn't need to be.
For example:
And then someone will replace regex with a wizard ("How many letters? What kind of letters? Should they be preceded by a space or followed by a tab? Do you want colons and double quotes with that?").
These are not mutually exclusive options. In fact, we would benefit a great deal from exposing to the user a tool which guided them through regex construction, and then showed them the regex as it was built up (and included say, tool-tips which would explain each part - I have such an app which I use for just this purpose).
I feel the "mutual exclusivity" is very much the crux of the matter though - the DE's we're all complaining about, lately, seem to be very much focused on putting these ideas in opposition.
The "simplified" interface should make way to the normal or advanced interface with an absolute minimum of effort - but I see nothing inherently wrong with trying to design the UI so that it slowly educates users about more advanced concepts in Linux/Unix (especially because, it's self-advertising as well - I've amazed people when I see them doing some repetitive task, and show them how a relatively simple shell script can automate it - but what would be even better, would be if we had a UI which got them into that mindset in the first place).
Pretty much agree with this post.
I'm always looking for a "best" way of doing something, but a lot of those choices would essentially be my "simplest" fallbacks when I want stuff to just work.
LXDE, interestingly, also has a much more "Unix" design philosophy behind it - self-contained apps which do a single-task very very well.
My first exposure to it was entirely without realizing when I was trying to figure out if I could put a touchscreen interface together using just "panel" apps that ran the appropriate commands for different tasks. The LX music player delivered.
IMO it represents a much more correct design philosophy for building a useful Linux desktop under open-source development, especially when you note that a lot of the current desktops that have pretty much gone to hell, went there around the same time they decided they were going to "integrate" everything.
The "four corners" idea gets implemented to the detriment of other ideas though. It's just not that important, and its probably the number 1 thing I despise about Mac OS (and by extension, Unity) - constantly swapping around my application menus at the top of the screen.
"Mean time to console" should probably be a UI metric - how long after installing a new DE does the average user find they need to open up a text configuration file or run a console command to fix or customize something.
Or you know, just don't do it.
Most of the forced paradigm shifts in recent years are notable because there's absolutely nothing about them which requires a forced paradigm shift. The Office "Ribbon" comes to mind (the entire UI framework easily supports traditional style toolbars, there's no reason a "2003 mode" couldn't have been left in).