I agree, cs needs a split in roles kinda like nurse and doctor.
A technician at CERN maintaining the accelerators is probably pretty damn smart, they probably also understand some particle physics. They may not be the ones driving/designing the experiments and they probably aren't getting their names on those physics papers but that doesn't mean they didn't contribute any value...
Which isn't the difference between doctor and nurse, more like an MD and PhD in medicine, one focused on practice, the other research.
In school terms it would be more like a technician or vocational training around coding for particular types of projects.
Computer Scientists - Come up with the foundational algorithms etc.
Software Engineers - Translate and map those foundational algorithms into more readily available tooling, occasionally directly building a complex system using engineering principles.
Software Developers - Use the tooling created by software engineers to create user friendly solutions for the masses, aka developing apps or websites.
Of course, no manager reads The Mythical Man Month, and those that do don’t follow it.
I've literally never met a manager that practiced this. Is the book still relevant? Not disagreeing, just curious.
Yeah it's a pretty universal idea.
> Is the book still relevant?
Yes, precisely because that idea is so universal. Its advice has been relevant, insightful, and mostly ignored ever since it was written.
The book is worthwhile, not just for the central lesson, but WHY (TL;DR: communication overhead between different individuals increase as you add "nodes" - interestingly, you can make analogies to L2 cache expiration problems and others) and also a nice view at IBM of the past. And it was written in an era where there wasn't the need to make every book 300 pages long, so it's generally a lot more meat/page, though hardly perfect.
Here's quite a famous example: https://www.statista.com/statistics/272140/employees-of-twit... It shows a graph of the number of employees at Twitter over time. In Jan '08, there are 8. In Jan '09 there are 29. In Jan '10 there are 130. In Jan '11 there are 350. In Dec '13 there are 2712.
Although a lot of that staff are going to be sales and marketing (and probably the growth is justified), the development team is also growing exponentially during that time. The Mythical Man Month would say that probably they are spending a lot more money than they need to to get the growth they needed.
I used to work for a manufacturer of telephony equipment. On one of the products I worked on they had 5000 developers! (A single piece of code!!!). The average amount of code deployed was 1 line of code per day per developer. As much as we can argue that KLOC is a bad measure of productivity, if your average developer is only producing considerably less than 1 KLOC in an entire year, you know you have really, terrible, terrible problems. One of the questions you might want to ask is, if you want to write about 5000 lines of code a day, where is the sweet spot, actually? I think we can agree it's not at 5000 programmers. However, it's often really, really difficult to talk to non-technical management and get them to understand that more programmers does not usually equal more productivity.
I could regale you with literally hundreds more examples, but I think it's sufficient to say that, yes, the Mythical Man Month is still really relevant these days.