3 ms·

The problem is not that it was easier in the 80s to start programming (that would be that same as to say it was easier to study medicine in the 1800s) but that

The problem is not that it was easier in the 80s to start programming (that would be that same as to say it was easier to study medicine in the 1800s) but that today it is much harder to get to the point to do something meaningful. If you want the simplicity of the Basic interpreter, just fire up a Python console, and you are in a much more comfortable position to learn programming and computer science than you were back then. But it is still a long way to get to someplace useful. In the 80s, by owning a VC-20 and programming Basic and Assembler, I was pretty much already at the edge of something new and powerful.

AnimalMuppet 8y ago
> The problem is not that it was easier in the 80s to start programming (that would be that same as to say it was easier to study medicine in the 1800s) but that today it is much harder to get to the point to do something meaningful. I think that in the 1980s, the threshold of meaningful was a lot lower. It wasn't just that you needed less of a stack to be able to do things; it's that so much less had already been done. And those things hadn't been done because they were too hard to do with the existing tools. But we suddenly had a new, more powerful set of tools, and we could do a bunch of things that couldn't be done before. (I remember, in high school, moving from punched cards, where it took a week to see the results of a change, to BASIC on a TRS-80, where it took a minute. That changed the world quite a bit...)
dragonwriter 8y ago
I think that still misses the point a bit, but is closer. The availability of more ready-to-use software means that even though it's much easier to do something meaningful/useful with the available programming tools, it's much harder to find an application where programming it yourself is the most efficient way to acheive that utility, so there is a lot more needed most of the time before approaching a problem with programming is a net benefit independent of developing programming skills.
autokad 8y ago
> "but that today it is much harder to get to the point to do something meaningful." in some ways that is true, in many ways that is false. only as far back as 2005~, if I didnt know linux and wanted to learn it, I needed to purchase hardware. installing and messing with binaries / configurations ran the risk of making a machine inoperable. now today, its very easy to spin up something in AWS or VM. want to experiment with tensorflow on a large GPU box? 4$/hour on demand. also, we have lots of API's and libraries we can just import, and I think thats a great thing. writing a website where the login was needed but not ultimately what you wanted to explore use to be hard. now its gem install devise. makinga basic website look pretty enough? bootstrap. want to compare random forests, xgboost, and linear regression for predicting home prices? from sklearn import ... i know all these things arent doing 'new' things, but they enable the person to not worry about the details and do something cool. I think the volume of interesting stuff we see coming from blogs and articles on hacker news is a testament to that.
- aidenn0 8y ago
  I think the point is that in 1983, if you wrote a program that read in some input, did some calculations and printed output, you had created a program that resembled much of the professional software out there. If you have kids do that know they would probably say "It's not a real computer program, it's just text!"

AnimalMuppet 8y ago

> The problem is not that it was easier in the 80s to start programming (that would be that same as to say it was easier to study medicine in the 1800s) but that today it is much harder to get to the point to do something meaningful. I think that in the 1980s, the threshold of meaningful was a lot lower. It wasn't just that you needed less of a stack to be able to do things; it's that so much less had already been done. And those things hadn't been done because they were too hard to do with the existing tools. But we suddenly had a new, more powerful set of tools, and we could do a bunch of things that couldn't be done before. (I remember, in high school, moving from punched cards, where it took a week to see the results of a change, to BASIC on a TRS-80, where it took a minute. That changed the world quite a bit...)

dragonwriter 8y ago

I think that still misses the point a bit, but is closer. The availability of more ready-to-use software means that even though it's much easier to do something meaningful/useful with the available programming tools, it's much harder to find an application where programming it yourself is the most efficient way to acheive that utility, so there is a lot more needed most of the time before approaching a problem with programming is a net benefit independent of developing programming skills.

autokad 8y ago

> "but that today it is much harder to get to the point to do something meaningful." in some ways that is true, in many ways that is false. only as far back as 2005~, if I didnt know linux and wanted to learn it, I needed to purchase hardware. installing and messing with binaries / configurations ran the risk of making a machine inoperable. now today, its very easy to spin up something in AWS or VM. want to experiment with tensorflow on a large GPU box? 4$/hour on demand. also, we have lots of API's and libraries we can just import, and I think thats a great thing. writing a website where the login was needed but not ultimately what you wanted to explore use to be hard. now its gem install devise. makinga basic website look pretty enough? bootstrap. want to compare random forests, xgboost, and linear regression for predicting home prices? from sklearn import ... i know all these things arent doing 'new' things, but they enable the person to not worry about the details and do something cool. I think the volume of interesting stuff we see coming from blogs and articles on hacker news is a testament to that.

aidenn0 8y ago
I think the point is that in 1983, if you wrote a program that read in some input, did some calculations and printed output, you had created a program that resembled much of the professional software out there. If you have kids do that know they would probably say "It's not a real computer program, it's just text!"