3 ms·

This exactly! I like to make the car analogy: I have no idea how my car works. I mean, I more or less understand the principles underlying the internal combust

by chias 10y ago

This exactly! I like to make the car analogy: I have no idea how my car works. I mean, I more or less understand the principles underlying the internal combustion engine, but I wouldn't be able to service one, much less assemble one. But I don't need to. Typically, the only indication made available to me that something is wrong is a single bit of information ("check engine light"), but that is enough. You don't have to be a "car person" to make effective use of a car. I get in, I go, and well over 99% of the time that's the end of the story. Compare this with computers. When something goes wrong, it's usually vital that you (or your users) relay the precise error message (and God help you if there isn't one). You generally have to be a "computer person" to some degree to make effective use of a computer. If you are unconvinced by this comparison, contrast how often your family asks you to perform [computer task] versus how often you would approach a mechanic family member to perform [car task]; Contrast how often you hear "I can't do this, I'm not a car person" versus "I can't do this, I'm not a computer person". I consider swaths of modern hackers who simply don't know about much of ASCII as evidence of babysteps towards computers maturing as a technology.

ryandrake 10y ago
> If you are unconvinced by this comparison, contrast how often your family asks you to perform [computer task] versus how often you would approach a mechanic family member to perform [car task]; Contrast how often you hear "I can't do this, I'm not a car person" versus "I can't do this, I'm not a computer person". I think this is more about social norms/conventions than anything. I would never ask a family member who happens to be a surgeon to remove my gall bladder for me, or a family member who happens to be a mechanic to replace my clutch over the weekend. But for some crazy reason, it's perfectly acceptable to ask your "computer person" family members to spend hours removing the 1200 malware infections you got by installing that cute puppy toolbar you downloaded. The complexity of the tasks doesn't have anything to do with it.
joshuata 10y ago
I understand your argument, but I would disagree with a few points: First, we use computers for so many more things than cars. The average user does really well with basic tasks like checking their email and simple word processing. This would be daily driving in your car analogy. Occasionally things blow up, but that isn't too different from a major problem with a car. However users are constantly trying new things with computers, new programs, websites, and tasks. Car-owners who are constantly trying new things with their cars have as many problems, if not more, than the average computer user. The difference is that the people who use the full range of their car's capabilities are deeply interested in their vehicles. Second, abstractions like the check engine light are far from perfect. How do you know whether the light signals imminent failure or a minor inconvenience? What additional information is needed for the mechanic to diagnose the problem? I recently chased down a problem in my car that caused the check engine light to come on with a code that was physically impossible. It took a few weeks of careful experimentation and instrumentation before I was able to figure out what it thought was going on. This was a case where I absolutely needed more than a cursory knowledge of how my car works. I also think that a hacker should be similar to a amateur mechanic: although their car might be fuel-injected, they have a cursory knowledge of how a carburetor works. They may have an automatic transmission, but they understand what a clutch is. Compare that to many developers who have never set foot outside their niche; They have never used a radically different programming language or a different OS. They've never taken the time to dig into the layers beneath the one they use. I would argue that is a weakness. How will you ever debug a problem when it inevitably occurs in the layers beneath you?
digi_owl 10y ago
> how often you would approach a mechanic family member to perform [car task]; I guess it depends on where one live, as i see that happen all the time (either family, friends, or neighbors).
awfgylbcxhrey 10y ago
I have no idea how my car works. I find this weird. As I proceeded through Comp sci in high school, going from Pascal to C to assembler, I was always troubled by my lack of understanding, "but why does it work?" That anxiety finally disappeared in college when I learned how logic gates are constructed and went through the exercise of implementing multiply as a series of logical operations. Similarly, I find it strange that someone would be comfortable driving a car without fairly deep knowledge of how it functions and how to repair it. I don't understand how you're not plagued with anxiety.
i336_ 10y ago
> I consider swaths of modern hackers who simply don't know about much of ASCII as evidence of babysteps towards computers maturing as a technology. Nobody else has argued on this point yet so I'll throw my 2¢ in. (An aside: I had to look up the codepoint for ¢ - 2A - so I could use it. I haven't memorized ASCII yet, let alone Unicode.) In my opinion, things like the first 31 characters of ASCII, line discipline, NIC PHY AUIs, the difference between RS-232 (point-to-point) vs RS-485 (current-loop), how to make your Classic Mac show a photo of the developer team (hit the Interrupt key then input "G 41D89A"), or how to play notes on period VT100s (set the keyboard repeat rate really high); we've moved into an era where Web stacks reveal hard-to-diagnose bugs in nearly-40-year-old runtimes (Erlang), Apple will give you $200k if you extract the Secure Boot ROM out of your iPhone (in one person's case via a bespoke tool that attached to the board and talked PCIe), the UEFI in Intel NUCs is such a close match for the open-source BSP that Intel releases that it's a lot easier than everyone would like for you to make UEFI modules that step on things that shouldn't be step-on-able and let you fall through holes into SMM (Ring -2), and few people care that sudo on macOS doesn't really give you root-level privileges anymore. We've just replaced all the old idiosyncrasies with a bunch of more modern idio[syn]crasies. You're right that technology has matured, but this has unfortunately meant that a lot of the innocence we took for granted has been lost. Things aren't an absolute disaster, but it's more political now, and we have to keep on our toes. Computers aren't universally somewhere we can go to to have fun; we have to work to find the fun now. (Also, I just did that thing I often do with forums - I expect my reply to appear at the end of the thread, so I go to click on the reply button at the end. But that would reply to the comment at the end of the thread, not yours. This is a problem endemic to forum UI and not a HN issue. Not all aspects of computers have matured yet, not by a long shot.)

by chias 10y ago

ryandrake 10y ago

> If you are unconvinced by this comparison, contrast how often your family asks you to perform [computer task] versus how often you would approach a mechanic family member to perform [car task]; Contrast how often you hear "I can't do this, I'm not a car person" versus "I can't do this, I'm not a computer person". I think this is more about social norms/conventions than anything. I would never ask a family member who happens to be a surgeon to remove my gall bladder for me, or a family member who happens to be a mechanic to replace my clutch over the weekend. But for some crazy reason, it's perfectly acceptable to ask your "computer person" family members to spend hours removing the 1200 malware infections you got by installing that cute puppy toolbar you downloaded. The complexity of the tasks doesn't have anything to do with it.

joshuata 10y ago

I understand your argument, but I would disagree with a few points: First, we use computers for so many more things than cars. The average user does really well with basic tasks like checking their email and simple word processing. This would be daily driving in your car analogy. Occasionally things blow up, but that isn't too different from a major problem with a car. However users are constantly trying new things with computers, new programs, websites, and tasks. Car-owners who are constantly trying new things with their cars have as many problems, if not more, than the average computer user. The difference is that the people who use the full range of their car's capabilities are deeply interested in their vehicles. Second, abstractions like the check engine light are far from perfect. How do you know whether the light signals imminent failure or a minor inconvenience? What additional information is needed for the mechanic to diagnose the problem? I recently chased down a problem in my car that caused the check engine light to come on with a code that was physically impossible. It took a few weeks of careful experimentation and instrumentation before I was able to figure out what it thought was going on. This was a case where I absolutely needed more than a cursory knowledge of how my car works. I also think that a hacker should be similar to a amateur mechanic: although their car might be fuel-injected, they have a cursory knowledge of how a carburetor works. They may have an automatic transmission, but they understand what a clutch is. Compare that to many developers who have never set foot outside their niche; They have never used a radically different programming language or a different OS. They've never taken the time to dig into the layers beneath the one they use. I would argue that is a weakness. How will you ever debug a problem when it inevitably occurs in the layers beneath you?

digi_owl 10y ago

> how often you would approach a mechanic family member to perform [car task]; I guess it depends on where one live, as i see that happen all the time (either family, friends, or neighbors).

awfgylbcxhrey 10y ago

I have no idea how my car works. I find this weird. As I proceeded through Comp sci in high school, going from Pascal to C to assembler, I was always troubled by my lack of understanding, "but why does it work?" That anxiety finally disappeared in college when I learned how logic gates are constructed and went through the exercise of implementing multiply as a series of logical operations. Similarly, I find it strange that someone would be comfortable driving a car without fairly deep knowledge of how it functions and how to repair it. I don't understand how you're not plagued with anxiety.

i336_ 10y ago

> I consider swaths of modern hackers who simply don't know about much of ASCII as evidence of babysteps towards computers maturing as a technology. Nobody else has argued on this point yet so I'll throw my 2¢ in. (An aside: I had to look up the codepoint for ¢ - 2A - so I could use it. I haven't memorized ASCII yet, let alone Unicode.) In my opinion, things like the first 31 characters of ASCII, line discipline, NIC PHY AUIs, the difference between RS-232 (point-to-point) vs RS-485 (current-loop), how to make your Classic Mac show a photo of the developer team (hit the Interrupt key then input "G 41D89A"), or how to play notes on period VT100s (set the keyboard repeat rate really high); we've moved into an era where Web stacks reveal hard-to-diagnose bugs in nearly-40-year-old runtimes (Erlang), Apple will give you $200k if you extract the Secure Boot ROM out of your iPhone (in one person's case via a bespoke tool that attached to the board and talked PCIe), the UEFI in Intel NUCs is such a close match for the open-source BSP that Intel releases that it's a lot easier than everyone would like for you to make UEFI modules that step on things that shouldn't be step-on-able and let you fall through holes into SMM (Ring -2), and few people care that sudo on macOS doesn't really give you root-level privileges anymore. We've just replaced all the old idiosyncrasies with a bunch of more modern idio[syn]crasies. You're right that technology has matured, but this has unfortunately meant that a lot of the innocence we took for granted has been lost. Things aren't an absolute disaster, but it's more political now, and we have to keep on our toes. Computers aren't universally somewhere we can go to to have fun; we have to work to find the fun now. (Also, I just did that thing I often do with forums - I expect my reply to appear at the end of the thread, so I go to click on the reply button at the end. But that would reply to the comment at the end of the thread, not yours. This is a problem endemic to forum UI and not a HN issue. Not all aspects of computers have matured yet, not by a long shot.)