[crispinb]

> Knowing X language isn't what makes a good developer, it's knowing some sufficiently multi-paradigm langauge ...

True enough as far as it goes. But in practical terms, as platforms have multiplied in number and increased in surface area, being a useful and/or marketable developer means in-depth knowledge of specific platforms. It's much harder to spin up quickly on a platform than a mere language. However well you know java, you're not going to be a useful Android programmer without a decent understanding of the gamut of Activities, Fragments, dexing, etc. Cocoa is far more the key to iOS development than is Swift (or Objective-C). Go/Javascript/Java (etc) knowledge doesn't get you more 1/10th of the way to productivity with AWS.

All this is making it increasingly difficult to be a convincing generalist.

[hardwaresofton]

I agree with that point, except for two minor things:

- mobile development in particular is the worst possible case of this. mobile platform companies are in the best case optimized only for their own case, and in the worst case maliciously trying to create developer walled gardens. However, I would point out that knowing one of these often helps in knowing the other (Android vs iOS) -- there are often 1:1 concepts that map over, because in the end what you want to do is very much the same, and the way the platforms have modeled how to do certain things is often similar.

- AWS is not really a programming language, even if you could CloudFormation templates, which surely will be turing complete if they aren't considered to be already. "Familiarity with aws" roughly equals "familiarity with deploying an application to managed remote servers", which is orthogonal (IMO) to pure application development, but is increasingly a required skill.

I do agree that being a convincing generalist is difficult, and I think this is why generic `Backend Developer` or `Software Engineer` positions often pay more than `Mobile developer` or `Frontend Developer` positions (at least in my experience they have). This is also why people try to hire for "smarts" in interviews rather than just testing language/syntax knowledge -- generalists make for better cogs ("are more fungible" would be the nicer way to say it I guess).

[crispinb]

You're entirely right insofar as (eg) an Android developer will learn iOS far faster than someone without mobile (or any other ui kit) experience. But the specifics are different enough such that it's hard to do both to a high level. Or perhaps I'm just speaking for my own limitations -- I do iOS & Android (native & a bit of react native), but I'm definitely less of an expert with either (or the various server things I also do) than I would be if I specialised.

I was offering AWS as an example of another platform rather than a language. If you're called on to (for example) implement a bunch of microservice type stuff, there's quite a bit to learn about RDS/dynamo/lambdas/IAM etc, and then all the quite complex deployment stuff to make it truly production ready.

I think in all these cases, the language is a relatively trivial component. I generally can get quite productive in a new language (that isn't greatly paradigmatically different, ie. allowing for exceptions like Haskell, Rust etc) in a couple of days to a week. A new platform? More like weeks to months.

[hardwaresofton]

> You're entirely right insofar as (eg) an Android developer will learn iOS far faster than someone without mobile (or any other ui kit) experience. But the specifics are different enough such that it's hard to do both to a high level. Or perhaps I'm just speaking for my own limitations -- I do iOS & Android (native & a bit of react native), but I'm definitely less of an expert with either (or the various server things I also do) than I would be if I specialised.

Yeah, that's definitely true pretty much why projects like react native and flutter exist. I personally have done toy projects on iOS and Android, but do not do it for my day job (and really have no desire to), so I stick to projects like react native and flutter and try to avoid mobile development like the plague.

> I was offering AWS as an example of another platform rather than a language. If you're called on to (for example) implement a bunch of microservice type stuff, there's quite a bit to learn about RDS/dynamo/lambdas/IAM etc, and then all the quite complex deployment stuff to make it truly production ready.

Yeah, absolutely -- funnily enough, I'm relatively sure that AWS wants it this way. AWS can be come so complex so fast, especially when they have many ways to do the same thing (often built upon each other) and it isn't necessarily immediately clear which is the best for your usecase. The additional management they give you for some services is minimal -- I'm thoroaughly convinced that most of the "value" offered by RDS is more-or-less thanks to tools like postgres being well-written/robust software.

I agree on the language being relatively trivial too -- maybe the importance of platforms ties into that -- Once the language difference becomes trivial the next big differentiator must indeed be platforms/the well-used structures that are built on the base language. I'll make sure to express that more clearly in the future

[crispinb]

> I'm relatively sure that AWS wants it this way

Oh yes. Capture is definitely the name of the game. For 'engineers' or 'software developers' to become instead 'iOS/Android/AWS developers' is good business for the corporations concerned (and ultimately, in my opinion, very bad for us).