> I disagree about your premise at multiple points, but I want to focus on just one point - In my experience, Functional Programming is a large enough of a technical reason to not choose Scala in a company of a reasonable size.
You say you disagree with my premise, yet you seem to be confirming the point I made. Your arguments revolve around difficulties with finding devs for a team and with training them to actually understand FP at a level that would allow them to untangle their algorithmic thinking from OOP. It's exactly what I said about the only proper reason to choose Java over Scala: team's lack of FP expertise. Other than that, my point still holds: type algebra and explicit nulls opt-in bring better quality software overall. And it does require FP expertise, not just awareness or vague familiarity with.
> otherwise your devs will go off the rails using Scala in ways that the team isn't prepared to learn
This also contributes to my initial point. But I'd like to know what are your examples of off-the-rails, as I've been to situations where a couple of folded traversals had been called similarly "too much" in a Scala codebase, but for someone writing Haskell it was just a regular application of a familiar pattern.
> Sure, you could hire someone who knew functional programming already and came from a background writing Haskell, but they were rare and usually commanded exorbitant salaries that made sure they were never in the equation to begin with.
This is a specific argument that could be made within your organisation. I don't know your budgets and the requested salaries by respective candidates, and you didn't clarify the level of expertise your teams were looking for, so I can't make an informed judgement about the merits of this argument and to what extent it could be applied onto the tech stack in general. I've been doing FP for a meaningful amount of time, mostly in Haskell, and the common theme over all these years has always been a visible shortage of FP positions compared to the number of capable candidates willing to compromise on their salary in exchange for an opportunity to use Haskell at their workplace. Many of them were instead compromising on Haskell in exchange for a regular Java-grade salary for writing Scala. And I've never heard of people commanding better salaries just for the fact of being able to write Scala instead of Java. Usually it was due to additional set of skills, like strong algorithmic training or extra years of business domain expertise, and these traits bring additional value to organisations, if they are able to utilize them within their teams.
> Especially when Java is not really a terrible choice in a large enough organization
It certainly isn't terrible, but it also is neither safer nor more productive. For a developer who has a good command of both paradigms, Scala would be a better choice on JVM.
You say you disagree with my premise, yet you seem to be confirming the point I made. Your arguments revolve around difficulties with finding devs for a team and with training them to actually understand FP at a level that would allow them to untangle their algorithmic thinking from OOP. It's exactly what I said about the only proper reason to choose Java over Scala: team's lack of FP expertise. Other than that, my point still holds: type algebra and explicit nulls opt-in bring better quality software overall. And it does require FP expertise, not just awareness or vague familiarity with.
> otherwise your devs will go off the rails using Scala in ways that the team isn't prepared to learn
This also contributes to my initial point. But I'd like to know what are your examples of off-the-rails, as I've been to situations where a couple of folded traversals had been called similarly "too much" in a Scala codebase, but for someone writing Haskell it was just a regular application of a familiar pattern.
> Sure, you could hire someone who knew functional programming already and came from a background writing Haskell, but they were rare and usually commanded exorbitant salaries that made sure they were never in the equation to begin with.
This is a specific argument that could be made within your organisation. I don't know your budgets and the requested salaries by respective candidates, and you didn't clarify the level of expertise your teams were looking for, so I can't make an informed judgement about the merits of this argument and to what extent it could be applied onto the tech stack in general. I've been doing FP for a meaningful amount of time, mostly in Haskell, and the common theme over all these years has always been a visible shortage of FP positions compared to the number of capable candidates willing to compromise on their salary in exchange for an opportunity to use Haskell at their workplace. Many of them were instead compromising on Haskell in exchange for a regular Java-grade salary for writing Scala. And I've never heard of people commanding better salaries just for the fact of being able to write Scala instead of Java. Usually it was due to additional set of skills, like strong algorithmic training or extra years of business domain expertise, and these traits bring additional value to organisations, if they are able to utilize them within their teams.
> Especially when Java is not really a terrible choice in a large enough organization
It certainly isn't terrible, but it also is neither safer nor more productive. For a developer who has a good command of both paradigms, Scala would be a better choice on JVM.