4 ms·

The form and function of Unix-based operating systems hasn't really changed much since the 70s, but even then the most significant investment in their developme

by avar 8y ago

The form and function of Unix-based operating systems hasn't really changed much since the 70s, but even then the most significant investment in their development (Linux et al) has been in the last 20-30 years. As a sibling comment pointed out[1] better than I could there's perception bias in thinking that just because something outwardly looks similar, that changing it hasn't been just as hard or harder than initially getting to that point from earlier technology that looked quite different. It was much harder to go from a 737 in the 60s to a modern 787 in the 2010s, than from something like the Douglas DC-3 in the 30s to a 737 in the 60s, even though the 737 and 787 outwardly look about the same, and quite different from the DC-3. 1. https://news.ycombinator.com/item?id=18223205 https://news.ycombinator.com/item?id=18223205

joe_the_user 8y ago
As a sibling comment pointed out... better than I could there's perception bias in thinking that just because something outwardly looks similar, that changing it hasn't been just as hard or harder than initially getting to that point from earlier technology that looked quite different. Sure, I think in general once the form and use of a given product or technology is set, much more labor is going to be invested in perfecting and polishing it than was originally invested in creating it. This is entirely logical - once X is perfected, we know X is going to sell (or be used) so money spent making the most desirable version of X is well-spent. Whereas if X is still in flux and you spend a lot of time and labor on small changes to X, you might find out some other version of X appears, making your investment nil. Here, there's also the factor that the perfection and polishing of a given technology can mean that it stands in the way of changes that would otherwise improve things. And network effect increases this effect as well. Edit: and remember, all this is talking about why revolutionary changes to products and technologies stop.
- avar 8y ago
  > [...]why revolutionary changes to products > and technologies stop[...] Yeah, but maybe we're just impressed by different things. I'm more impressed by the number of air passengers carried yearly around the globe being equivalent to around 50% of the world population, v.s. around 8% in the 1970s[1]. I'm more impressed by aircraft being reliable enough that most modern twin-engine jets in service can get an ETOPS[2] rating. There's been an amazing growth in the amount of direct flights, both transoceanic and otherwise. It's like how nothing fundamentally new really happened in E-Mail since the 1970s, but in terms of making it a useful consumer technology iteration was everything. 1. https://data.worldbank.org/indicator/IS.AIR.PSGR https://data.worldbank.org/indicator/IS.AIR.PSGR 2. https://en.wikipedia.org/wiki/ETOPS https://en.wikipedia.org/wiki/ETOPS
  - xg15 8y ago
    > I'm more impressed by the number of air passengers carried yearly around the globe being equivalent to around 50% of the world population, v.s. around 8% in the 1970s[1]. This sounds more like innovation in how to run an airline than technological innovation though.
    - avar 8y ago
      No, it's mainly technological. Here you can see the breakdown of operating costs of different wide-body aircraft: https://www.planestats.com/bhsw_2014sep https://www.planestats.com/bhsw_2014sep As shown there majority of the cost is fuel, and as hcarvalhoalves's comment notes fuel economy has improved by 45% since the 60s. Not having to operate a 4-engine jumbo jet for transatlantic flights or make multiple landings in Ireland/Iceland/Nova Scotia just to fly between Paris & New York is also a huge reduction in cost. That's what ETOPS in my comment is referring to. The operating of having more engines is also higher. Flights that today cost $400 would cost something like $1000 today if it wasn't for these sorts of technological improvements, and $1000 flights would be $2500 etc.
rayiner 8y ago
> The form and function of Unix-based operating systems hasn't really changed much since the 70s, but even then the most significant investment in their development (Linux et al) has been in the last 20-30 years. Investment is of course different than results. New generations of chip fabrication are exponentially more expensive, but advances in processor speeds (especially for single threaded code) have plateaued.

by avar 8y ago

joe_the_user 8y ago

As a sibling comment pointed out... better than I could there's perception bias in thinking that just because something outwardly looks similar, that changing it hasn't been just as hard or harder than initially getting to that point from earlier technology that looked quite different. Sure, I think in general once the form and use of a given product or technology is set, much more labor is going to be invested in perfecting and polishing it than was originally invested in creating it. This is entirely logical - once X is perfected, we know X is going to sell (or be used) so money spent making the most desirable version of X is well-spent. Whereas if X is still in flux and you spend a lot of time and labor on small changes to X, you might find out some other version of X appears, making your investment nil. Here, there's also the factor that the perfection and polishing of a given technology can mean that it stands in the way of changes that would otherwise improve things. And network effect increases this effect as well. Edit: and remember, all this is talking about why revolutionary changes to products and technologies stop.

avar 8y ago
> [...]why revolutionary changes to products > and technologies stop[...] Yeah, but maybe we're just impressed by different things. I'm more impressed by the number of air passengers carried yearly around the globe being equivalent to around 50% of the world population, v.s. around 8% in the 1970s[1]. I'm more impressed by aircraft being reliable enough that most modern twin-engine jets in service can get an ETOPS[2] rating. There's been an amazing growth in the amount of direct flights, both transoceanic and otherwise. It's like how nothing fundamentally new really happened in E-Mail since the 1970s, but in terms of making it a useful consumer technology iteration was everything. 1. https://data.worldbank.org/indicator/IS.AIR.PSGR https://data.worldbank.org/indicator/IS.AIR.PSGR 2. https://en.wikipedia.org/wiki/ETOPS https://en.wikipedia.org/wiki/ETOPS
- xg15 8y ago
  > I'm more impressed by the number of air passengers carried yearly around the globe being equivalent to around 50% of the world population, v.s. around 8% in the 1970s[1]. This sounds more like innovation in how to run an airline than technological innovation though.
  - avar 8y ago
    No, it's mainly technological. Here you can see the breakdown of operating costs of different wide-body aircraft: https://www.planestats.com/bhsw_2014sep https://www.planestats.com/bhsw_2014sep As shown there majority of the cost is fuel, and as hcarvalhoalves's comment notes fuel economy has improved by 45% since the 60s. Not having to operate a 4-engine jumbo jet for transatlantic flights or make multiple landings in Ireland/Iceland/Nova Scotia just to fly between Paris & New York is also a huge reduction in cost. That's what ETOPS in my comment is referring to. The operating of having more engines is also higher. Flights that today cost $400 would cost something like $1000 today if it wasn't for these sorts of technological improvements, and $1000 flights would be $2500 etc.

rayiner 8y ago

> The form and function of Unix-based operating systems hasn't really changed much since the 70s, but even then the most significant investment in their development (Linux et al) has been in the last 20-30 years. Investment is of course different than results. New generations of chip fabrication are exponentially more expensive, but advances in processor speeds (especially for single threaded code) have plateaued.