4 ms·

I still find that Scala devs think they’re writing Haskell. From what I’ve seen in the community, no one cares about shipping code, they care about whether or n

I still find that Scala devs think they’re writing Haskell. From what I’ve seen in the community, no one cares about shipping code, they care about whether or not it’s “pure”.

thinkharderdev 6y ago
To give them their credit. The argument from the purists is that purity is itself about shipping code. As a long-time scala developer who has started using the pure-FP ecosystem (cats,cats-effect mainly) I am starting to see their point. It takes some time to grok but once you do then it is amazingly productive to work with.
- randmeerkat 6y ago
  Maybe you’re on an exceptional team, but I’ve never seen a Scala team that could deliver faster than a Python / Ruby / Java team. Scala teams also have a hard time hiring and if they hire a dev that’s new to Scala they then have to deal with a very long ramp up time.
  - thinkharderdev 6y ago
    Maybe. I certainly don't have data on it but am also not sure what sort of data would actually prove that point. Teams that choose Scala may just be solving different sorts of problems. Or maybe they take longer and deliver a higher quality product. All I know if that I am personally way more productive in Scala than any other language I've used (and I have used most of them at one point or another).
    - randmeerkat 6y ago
      > Or maybe they take longer and deliver a higher quality product. Without going into how do you define a higher quality product, does quality really matter that much..? Most code is going to be re-written every few years anyway. Software engineers aren’t building architectural wonders that will last hundreds or maybe thousands of years. So is the complexity of a language like Scala really worth it, just to deliver a product that will be obsolete in a few years anyway?
      - esarbe 6y ago
        Well, I've been working in financial and data services for many, many years now and I can tell you as much: many of the code bases I've been dealing with (and am dealing with) written in less strictly typed languages (much Python, lots and lots a lot Java) are riddled with hidden runtime errors that you'll just never see in Scala. If there's a bug, it's usually some modelling problem or some conceptual issues, not the implementation of the code. I've much more confidence in the code I've written in Scala than the code I've written in Python or Java and I never ever had to get up and pick up the phone in the middle of the night because some production issues with code written Scala.
        kaba0 6y ago
        Java has really great tooling to make it really safe if one wants to — the problem is usually the managements’ cheapness with programmers fresh out of school/bootcamp. There are areas with high security requirements where Java is used with great success.
      - thinkharderdev 6y ago
        On the margin, yes. If a language helps to deliver a better product (fewer bugs, more dependable, etc) then your prior should be that that is a good thing that will, all else equal, make for a better product. But I don't accept the premise that it is easier and faster to deliver a working product in a different language. Maybe it is in some cases but me and my time are most productive and produce the highest quality product using Scala. And I think that has a lot to do with the language and the ecosystem itself. But ymmv and if it's not a good fit for your team then you shouldn't use it.
        randmeerkat 6y ago
        I would make the counter argument that I would rather have a language that _may_ be more prone to bugs and be able to easily hire engineers. Also, there’s a lot of hidden costs in over engineering something. Trying to get something to be bug free is like trying to add another 9 to five 9s of reliability, the cost is rarely justified.

I still find that Scala devs think they’re writing Haskell. From what I’ve seen in the community, no one cares about shipping code, they care about whether or not it’s “pure”.

thinkharderdev 6y ago

To give them their credit. The argument from the purists is that purity is itself about shipping code. As a long-time scala developer who has started using the pure-FP ecosystem (cats,cats-effect mainly) I am starting to see their point. It takes some time to grok but once you do then it is amazingly productive to work with.

randmeerkat 6y ago
Maybe you’re on an exceptional team, but I’ve never seen a Scala team that could deliver faster than a Python / Ruby / Java team. Scala teams also have a hard time hiring and if they hire a dev that’s new to Scala they then have to deal with a very long ramp up time.
- thinkharderdev 6y ago
  Maybe. I certainly don't have data on it but am also not sure what sort of data would actually prove that point. Teams that choose Scala may just be solving different sorts of problems. Or maybe they take longer and deliver a higher quality product. All I know if that I am personally way more productive in Scala than any other language I've used (and I have used most of them at one point or another).
  - randmeerkat 6y ago
    > Or maybe they take longer and deliver a higher quality product. Without going into how do you define a higher quality product, does quality really matter that much..? Most code is going to be re-written every few years anyway. Software engineers aren’t building architectural wonders that will last hundreds or maybe thousands of years. So is the complexity of a language like Scala really worth it, just to deliver a product that will be obsolete in a few years anyway?
    - esarbe 6y ago
      Well, I've been working in financial and data services for many, many years now and I can tell you as much: many of the code bases I've been dealing with (and am dealing with) written in less strictly typed languages (much Python, lots and lots a lot Java) are riddled with hidden runtime errors that you'll just never see in Scala. If there's a bug, it's usually some modelling problem or some conceptual issues, not the implementation of the code. I've much more confidence in the code I've written in Scala than the code I've written in Python or Java and I never ever had to get up and pick up the phone in the middle of the night because some production issues with code written Scala.
      - kaba0 6y ago
        Java has really great tooling to make it really safe if one wants to — the problem is usually the managements’ cheapness with programmers fresh out of school/bootcamp. There are areas with high security requirements where Java is used with great success.
    - thinkharderdev 6y ago
      On the margin, yes. If a language helps to deliver a better product (fewer bugs, more dependable, etc) then your prior should be that that is a good thing that will, all else equal, make for a better product. But I don't accept the premise that it is easier and faster to deliver a working product in a different language. Maybe it is in some cases but me and my time are most productive and produce the highest quality product using Scala. And I think that has a lot to do with the language and the ecosystem itself. But ymmv and if it's not a good fit for your team then you shouldn't use it.
      - randmeerkat 6y ago
        I would make the counter argument that I would rather have a language that _may_ be more prone to bugs and be able to easily hire engineers. Also, there’s a lot of hidden costs in over engineering something. Trying to get something to be bug free is like trying to add another 9 to five 9s of reliability, the cost is rarely justified.