5 ms·

I think that this is good because it really does describe what software dev used to be like, and remnants of this still exist in a lot of major institutions. I'

I think that this is good because it really does describe what software dev used to be like, and remnants of this still exist in a lot of major institutions. I've recently had to jam a cloud first solution through an architecture review board with largely hands off keyboard type jobs, so the priests at least rang true to me. It's easy to overestimate how hard open source has won, and how little influence "Not invented here complex" has on most systems. Most software now comes in a stack, and other companies use the same stack with some light add ons. The most common experience of software development wasn't like that as late as 2010. The New York times used to maintain it's own programming language whose only application was the NYT website. Trello did too, and there are many companies that did that without hitting the conference circuit about it. The big shift that made the world we're in now is the idea of domain specific languages largely going away and the idea of software being an integrative practice replacing them. The big cultural marker might be MIT switching their intro course from scheme to python in the late 00s.

pydry 4y ago
Used to be like? IME NIH syndrome is alive and well. Not that this article would be entirely off base even if it wasnt.
- awkward 4y ago
  There's always going to be that guy. The question is whether he builds his own shed or gets to design the whole skyscraper.
  - g0nk2 4y ago
    The equation is better expressed as "Build versus Buy", imho. Actually it seems to me to be a 'natural law' of software development - no matter the discipline/industry/inkling - that on certain event horizons, one must decide whether to build something, or buy it. For many people, the act itself is all that matters. For some, its the users who decide.
    - awkward 4y ago
      I think you might be that guy. Don't worry about it - that guy gets to work on some cool stuff.
908B64B197 4y ago
> The New York times used to maintain it's own programming language whose only application was the NYT website. Trello did too, and there are many companies that did that without hitting the conference circuit about it. C#, Swift and to a lesser degree Hack are still very much alive. But the two former are public by design. There's no advantages anymore to having a custom DSL. > The big cultural marker might be MIT switching their intro course from scheme to python in the late 00s. The target audience of the course also changed quite a bit. Now it's taken by undergrads of all majors while it used to be taught to a mostly CS audience. With that in mind, it makes sense to teach Python, since it's nearly universal and has nice data-analysis libraries that are well documented.

pydry 4y ago

Used to be like? IME NIH syndrome is alive and well. Not that this article would be entirely off base even if it wasnt.

awkward 4y ago
There's always going to be that guy. The question is whether he builds his own shed or gets to design the whole skyscraper.
- g0nk2 4y ago
  The equation is better expressed as "Build versus Buy", imho. Actually it seems to me to be a 'natural law' of software development - no matter the discipline/industry/inkling - that on certain event horizons, one must decide whether to build something, or buy it. For many people, the act itself is all that matters. For some, its the users who decide.
  - awkward 4y ago
    I think you might be that guy. Don't worry about it - that guy gets to work on some cool stuff.

908B64B197 4y ago

> The New York times used to maintain it's own programming language whose only application was the NYT website. Trello did too, and there are many companies that did that without hitting the conference circuit about it. C#, Swift and to a lesser degree Hack are still very much alive. But the two former are public by design. There's no advantages anymore to having a custom DSL. > The big cultural marker might be MIT switching their intro course from scheme to python in the late 00s. The target audience of the course also changed quite a bit. Now it's taken by undergrads of all majors while it used to be taught to a mostly CS audience. With that in mind, it makes sense to teach Python, since it's nearly universal and has nice data-analysis libraries that are well documented.