[1]: https://en.wikipedia.org/wiki/Usage_share_of_operating_syste...
Again, I'm not talking about what's more powerful, and I'm not saying that everyone works this way, but I am suggesting the majority of developers do need the command line to do their job.
Unfortunately, I'm not aware of any usage statistics for the command line in isolation. But if you look at the most popular technologies, e.g., Node[2], then you can most likely extrapolate that the majority of developers are working with the command line.
[2]: https://insights.stackoverflow.com/survey/2019#technology-_-...
Just like we used to live on MPW and Metrowerks before.
Scripting is perfectly doable from macros and IDE callable scripts.
Again, I wasn't making argument about where developers spend most of their time. Just that they need the command line as part of their workflow.
The large majority of Apple and Microsoft developer communities live on their IDEs as part of our workflow.
Scripting and automatization can be easily done from the IDE as well, thanks to macros, REPL and build integration points.
Dropping down into a UNIX like CLI is the exception, not the norm for a large community.
Ah, but what about WSL? Well, Microsoft saw a market opportunity to win developers that buy Apple computers to do Linux work, instead of buying laptops from the likes of Dell/Asus/Tuxedo/system76 to start with.
So they are making UNIX devs comfortable on a foreign platform, just like NeXT was built on top of UNIX to take a piece of the pie from the UNIX workstation market being lead by Sun.
Using the shell was never a thing for NeXTSTEP nor classical Mac OS development workflow.
Regarding Xcode being implemented by managing processes, my point is that IDE's and text editors have moved to a model of using external processes to implement features. E.g., language server protocol, linters. That iOS bans this type of application is why iOS for programming is a wasteland. The tragedy of that fact is my whole point in this thread.
For example I can do everything I need from Python or PowerShell from inside Visual Studio, just like Xerox PARC devs used to do with their REPLs on Lisp, Smalltalk and Mesa/Cedar workstations.
Naturally I am forced to drop down to old style command line every now and then, but that is forced upon me by tools I don't create, and most of the time they are ports from UNIX tooling.
The question I pose then, is if these things are possible to do in other ways, and those other ways are better, why don't professional developers choose to use them? For example, why isn’t programming on an iPad popular? I.e., where these other methods are literally the only way to do programming on device? Using SSH to program entirely in a terminal is almost certainly more popular for professional developers than any other approach of programming on an iPad (using blog posts about how professional developers develop on an iPad as an indicator).
This is how Apple devs actually do development on iPads.
- Swift Playgrounds
- Pythonista
- Continuous for C# and F#
- Lisping
- Raskell
- GLSL Studio
Same applies to developers on Amiga and AtariST, also less professional?
Or what about Windows developers, that hardly bothered with cmd.com, doing their scripting via VBScript/JSCript and everything else via Visual Studio, as it was common practice until PowerShell, with its PowerShell IDE sprung into existence, also a bunch of amateurs?
Don't judge professional developers on other platforms with UNIX glasses on.
Not everyone is writing UNIX daemons, stuck in UNIX workflows.
As I said, a graphical REPL for scripting languages is much more powerful.