I agree with your point overall but disagree with one point:
>Your co-worker's path is dangerous. Replace "Microsoft" with "COBOL" or "Fortran" or "Mainframe" and you'll see that this is a problem.
The coworker is described as being "in his fifties." It's probably not unreasonable at that point to have significant depth in a single ecosystem. COBOL was exactly the case I was thinking of: very few (if any) new COBOL programmers are being minted any more and there is still a lot of legacy code. It will probably be a good, if dull, career for the next 20 years for that coworker who will at the end be able to write his own ticket to fit around his almost-retired status.
For someone with 40-50 years of working life ahead of him/her, this sort of specialization is indeed premature and dangerous.
For me that kind of specialization has never been attractive, an attitude that has served me well. But that's not true for everyone, at every stage of life.
To be fair, COBOL stopped evolving a long time ago, while .NET hasn't. He's in his fifties, but he's not working on legacy code; he grew up with the tip of Microsoft's technology. For example, he has pet projects on Azure and just completed a project at work in .NET 4.5 and Entity Framework 6. Sure, it's still a lock-in, and potentially dangerous, but nowhere near as COBOL or Fortran.
Yep, you've got it. So my point was directed against pre/chosen specialization for more junior engineers.
I should say that some level of specialization over time is natural. At some point you'll likely work on one thing for an extended period of time, and you'll build a bunch of domain knowledge.
In my case, I'm fairly biased against developers who frame their skills in relation to a specific technology - .NET Software Developer, Java Software Developer, iOS Software Developer. I really want people who think of themselves as Software Developers first and don't worry too much about the stack/domain.
The coworker is described as being "in his fifties." It's probably not unreasonable at that point to have significant depth in a single ecosystem. COBOL was exactly the case I was thinking of: very few (if any) new COBOL programmers are being minted any more and there is still a lot of legacy code. It will probably be a good, if dull, career for the next 20 years for that coworker who will at the end be able to write his own ticket to fit around his almost-retired status.
For someone with 40-50 years of working life ahead of him/her, this sort of specialization is indeed premature and dangerous.
For me that kind of specialization has never been attractive, an attitude that has served me well. But that's not true for everyone, at every stage of life.