6 ms·

This might have been true 20 years ago when software development meant everything from a fundamental understanding of hardware (including how pixels are laid ou

by zanny 8y ago

This might have been true 20 years ago when software development meant everything from a fundamental understanding of hardware (including how pixels are laid out, how to manage gamma, etc for the front end) to knowing a lot of theory (because you had to implement your collections in house, manage the code base by hand, etc) but nowadays I'm pretty sure at least half of everyone could be pointed at a web framework and bootstrap and make a functional website in half a year. Sure a professional would have made that in a week, but their next site will take half as long, and then half as long... until they are within striking distance of someone who does still know all those fundamentals. They will mostly be limited in what they can do, but most people don't really need developers that can do anything even if they pay for them without realizing that. Programming is being commoditized. In the same way you don't see a doctor first thing when you go to one; you see a nurse, tech, or assistant. Not everything requires the most expensive learning investment. Most programming jobs don't require you to be able to write a proof of the code you copy paste from Stack Overflow.

shados 8y ago
Very true. The most important thing you mentioned is this one IMO: > even if they pay for them without realizing that In health care, you have doctors, physician assistants, nurses, administrators, etc. They have very different background, schooling and certifications as well as salaries. In software, you have some differences (individual contributors, tech leads, managers, etc), but it's not granular at all within the categories. While the self taught html dev is making 90-100k, the MIT graduate builting complex systems with 5 years of experience is making 150k (Using round numbers here, don't read too much into it), when the difference is significantly bigger. People are quick to point out at flukes: there's totally exceptions, the CMU grad who can't code themselves out of paper bag, and the random highschool dropout who has been coding since they were 7 who are now world famous geniuses...but they're exceptions yet treated as the rule.
- therealdrag0 8y ago
  Agreed. I've seen people trained as other engineers (EE,ME) who have taken one programming class, get jobs at similar salaries as CS grads, but then need a long time to ramp up and are a "burden" to their coworkers who have to train them while also doing their own work. It's strange how much leeway people can get in this field. I understand anyone can be trained, but still is a luxurious field.
ntsplnkv2 8y ago
Programming has already been commoditized. No one has to learn anything. Simply outsource the code to a shop in India. Done. It's easy for those small problems. It's hard when you're designing products that need to get to market quickly and solve a difficult problem without product sinking bugs. That's why developers and engineers get paid a lot more.
kojeovo 8y ago
It's easy to make a website look pretty - which is what I find bootcamp people can be good at. That's a very small part of software development. They've, in my experience, struggled with highly distributed and scalable applications/systems.
- sombremesa 8y ago
  Most people don't need to make their own solutions for that anymore either, since we have AWS, GCP, GKE, etc. However, we software engineers (as well as computer scientists and devs) tend to underestimate how much we've already learned and internalized. It becomes painfully obvious when you try to teach someone, and realize that you need to start at a much earlier point than you expected.
  - mrep 8y ago
    What's funny is that I have only worked for the big tech companies since I graduated and I know how to do all the backend stuff really well and thus I can do it in a few weeks for my side projects. However, the fronted code kills me and I keep stalling and getting confused on my new projects trying to figure out all the js/html/css to make it look really good for normal users...
    - AlexCoventry 8y ago
      Just make it functional with react/redux, then hire a designer to give you feedback on it.
erikpukinskis 8y ago
Anyone can build an app, just as anyone can build a house. But if you want a roof that won’t rot for 20 years, you need to pay someone who knows how to build a proper house. By the same token, if you want an app that will still be maintainable in 5 years, you need a professional. An amateur can build most requirements, but the architecture will seize up if it’s just layers of hacks. It’s sort of an interesting situation for startups, because their business requirements for newly acquired users are often changing on <5 year timeframes. So actually throwing away the whole system every few years turns out to be an OK strategy. The problem is previously acquired customers, who won’t generally deal well if you tell them half of your products features are going away for 18 months, or maybe forever. You can sometimes hack around this by basically ditching all of your old customers and taking a high turnover strategy. That can work well for the middle of your growth curve. You burn a chunk of your addressable market in the long term. Also most companies don’t have the nerve to do it. And it can cause cash problems. But there are viable strategies that way.
- shados 8y ago
  > if you want an app that will still be maintainable in 5 years That's part of the issue. People have been making teams who CANT build apps still maintainable in 5 years. Everyone has and only a select few remember/know that it's possible. At a previous job, I was in a meeting where we pointed out an app had become unmaintainable to the point of requiring a rewrite. One of the engineers who built it pointed out "Well, we built this app a year and a half ago, of course it's unmaintainable, what do you expect?". I choked on my drink.
- alexhutcheson 8y ago
  Eh, I haven’t seen any evidence that experienced professionals can build a website that will last for 5 years either. Business requirements inevitably requires hacks and kludges, and abstraction layers implemented to deal with the problems of today become tomorrow’s layer that everyone wants to rewrite or refactor.
  - shados 8y ago
    > I haven’t seen any evidence that experienced professionals can build a website that will last for 5 years either You need a team of experienced professionals, which virtually no longer exists. You could work at 10 different companies back to back and not come across a team that can do it. That's the industry now.

by zanny 8y ago

shados 8y ago

Very true. The most important thing you mentioned is this one IMO: > even if they pay for them without realizing that In health care, you have doctors, physician assistants, nurses, administrators, etc. They have very different background, schooling and certifications as well as salaries. In software, you have some differences (individual contributors, tech leads, managers, etc), but it's not granular at all within the categories. While the self taught html dev is making 90-100k, the MIT graduate builting complex systems with 5 years of experience is making 150k (Using round numbers here, don't read too much into it), when the difference is significantly bigger. People are quick to point out at flukes: there's totally exceptions, the CMU grad who can't code themselves out of paper bag, and the random highschool dropout who has been coding since they were 7 who are now world famous geniuses...but they're exceptions yet treated as the rule.

therealdrag0 8y ago
Agreed. I've seen people trained as other engineers (EE,ME) who have taken one programming class, get jobs at similar salaries as CS grads, but then need a long time to ramp up and are a "burden" to their coworkers who have to train them while also doing their own work. It's strange how much leeway people can get in this field. I understand anyone can be trained, but still is a luxurious field.

ntsplnkv2 8y ago

Programming has already been commoditized. No one has to learn anything. Simply outsource the code to a shop in India. Done. It's easy for those small problems. It's hard when you're designing products that need to get to market quickly and solve a difficult problem without product sinking bugs. That's why developers and engineers get paid a lot more.

kojeovo 8y ago

It's easy to make a website look pretty - which is what I find bootcamp people can be good at. That's a very small part of software development. They've, in my experience, struggled with highly distributed and scalable applications/systems.

sombremesa 8y ago
Most people don't need to make their own solutions for that anymore either, since we have AWS, GCP, GKE, etc. However, we software engineers (as well as computer scientists and devs) tend to underestimate how much we've already learned and internalized. It becomes painfully obvious when you try to teach someone, and realize that you need to start at a much earlier point than you expected.
- mrep 8y ago
  What's funny is that I have only worked for the big tech companies since I graduated and I know how to do all the backend stuff really well and thus I can do it in a few weeks for my side projects. However, the fronted code kills me and I keep stalling and getting confused on my new projects trying to figure out all the js/html/css to make it look really good for normal users...
  - AlexCoventry 8y ago
    Just make it functional with react/redux, then hire a designer to give you feedback on it.

erikpukinskis 8y ago

Anyone can build an app, just as anyone can build a house. But if you want a roof that won’t rot for 20 years, you need to pay someone who knows how to build a proper house. By the same token, if you want an app that will still be maintainable in 5 years, you need a professional. An amateur can build most requirements, but the architecture will seize up if it’s just layers of hacks. It’s sort of an interesting situation for startups, because their business requirements for newly acquired users are often changing on <5 year timeframes. So actually throwing away the whole system every few years turns out to be an OK strategy. The problem is previously acquired customers, who won’t generally deal well if you tell them half of your products features are going away for 18 months, or maybe forever. You can sometimes hack around this by basically ditching all of your old customers and taking a high turnover strategy. That can work well for the middle of your growth curve. You burn a chunk of your addressable market in the long term. Also most companies don’t have the nerve to do it. And it can cause cash problems. But there are viable strategies that way.

shados 8y ago
> if you want an app that will still be maintainable in 5 years That's part of the issue. People have been making teams who CANT build apps still maintainable in 5 years. Everyone has and only a select few remember/know that it's possible. At a previous job, I was in a meeting where we pointed out an app had become unmaintainable to the point of requiring a rewrite. One of the engineers who built it pointed out "Well, we built this app a year and a half ago, of course it's unmaintainable, what do you expect?". I choked on my drink.
alexhutcheson 8y ago
Eh, I haven’t seen any evidence that experienced professionals can build a website that will last for 5 years either. Business requirements inevitably requires hacks and kludges, and abstraction layers implemented to deal with the problems of today become tomorrow’s layer that everyone wants to rewrite or refactor.
- shados 8y ago
  > I haven’t seen any evidence that experienced professionals can build a website that will last for 5 years either You need a team of experienced professionals, which virtually no longer exists. You could work at 10 different companies back to back and not come across a team that can do it. That's the industry now.