With any other tool, there would eventually be an onboarding process not just for the thing I built, but also for the tool I built it in. Not with Excel – that will just run, whereever you go.
With any other tool, there would eventually be an onboarding process not just for the thing I built, but also for the tool I built it in. Not with Excel – that will just run, whereever you go.
How to change it, yes. How to get the correct result, probably not.
But well, they will get a result. Many people are satisfied with that.
I'm not sure a receptive t will be able to successfully edit a spread sheet. So easy, to get interdependency between cells to a degree that it is really hard to follow, if you didn't create it yourself.
Or did I just describe Perl? Not sure.
No. That is not Perl. That is C Recursion: https://www.bobhobbs.com/files/kr_lovecraft.html
Python3 is up and coming, and eventually it will get there, but old machines must die before it can reach ubiquity.
awk, bc, and sh are POSIX base utilities, they are literally preinstalled on every unix derivative.
That doesn't make them linga franca, which is what Scarbutt was asking about.
Hell, ed is a POSIX base utility. Does that make ed's command mode the linga franca of text manipulation? Of course not, the thought is preposterous.
That it is an advantage does not naturally and unfailingly translate into being a linga Franca.
"vi" being so common (ie. "lingua franca") is what gave it an advantage.
I took it to mean "so common that it's available almost anywhere" (which is consistent with the context of the comment).
I've found it extremely useful to rely in tools such as the ones you just listed. Specially because they are often present even in light vms were perl might not.
Yes, I would consider awk and sh lingua francae (?) alongside Perl – but in slightly different contexts, namely low-complexity jobs.
I find Perl slightly more suitable once things get a little complicated, mainly because it's footgun:power ratio is better than awk (few footguns, but also very low power) and sh (many footguns, not that powerful).
I have come across plenty of boxen where bc was not installed, so I don't put that in that bucket.
And yes, if I want to describe text changes in a highly portable way, I will use sed (or ed, depending on context) rather than, say, unified diffs. I could accept an argument that unified diffs are also a lingua franca for more complicated changes, but they are harder to write.
Python breaks backwards compatibility with old scripts every six months or so. In practice, this means you have to port scripts every time you switch machines. Virtual environments sometimes address this issue, but they are hit or miss.
In contrast, perl scripts from 1999 often work unchanged on clean 2022 OS installs.