3 ms·

If anything is going to displace bash at this point it's going to be something that brings substantial benefits to the user. Stricter syntax is good but not of

by zimbatm 11y ago

If anything is going to displace bash at this point it's going to be something that brings substantial benefits to the user. Stricter syntax is good but not of the immediate concern to the user typing his 2x-longer command. The next evolution of shell is going to be something different. Something that doesn't depend on an antiquated TTY protocol, something that can display images and other constructs more complex that text, and more importantly something that allows the user to accumulate commands instead of typing the same thing over and over again. Maybe something like ipython/jupyter

pekk 11y ago
Using languages other than shell already brings lots of benefits to the user. That just doesn't matter to the people who are writing large programs in shell. It isn't like those people are basing their decision on rigorous user studies.
Aleman360 11y ago
PowerShell
hyperpape 11y ago
There's the shell language and the shell interface. The design of those things influence each other, but they're fundamentally separate even though bash/zsh/fish often make changes to both. The original article is really about languages, and I think that even with the best interface in the world, we want a better language.
- zimbatm 11y ago
  Just changing the language is in the diminishing return area and is why it hasn't happened yet. Something else need to change. Any sort of shell programming is going to be full of side-effects. No matter how strict the language, building the script will still involve trying on the target system, maybe accumulating shell commands in a script, and then have these long full cycles where the script is tested on a fresh system to make sure it's still working. This is happening every time I do any sort of sysadmin-type of stuff. Even in ansible, chef, docker, ... To escape that something deeper needs to be changed.

by zimbatm 11y ago

pekk 11y ago

Using languages other than shell already brings lots of benefits to the user. That just doesn't matter to the people who are writing large programs in shell. It isn't like those people are basing their decision on rigorous user studies.

Aleman360 11y ago

PowerShell

hyperpape 11y ago

There's the shell language and the shell interface. The design of those things influence each other, but they're fundamentally separate even though bash/zsh/fish often make changes to both. The original article is really about languages, and I think that even with the best interface in the world, we want a better language.

zimbatm 11y ago
Just changing the language is in the diminishing return area and is why it hasn't happened yet. Something else need to change. Any sort of shell programming is going to be full of side-effects. No matter how strict the language, building the script will still involve trying on the target system, maybe accumulating shell commands in a script, and then have these long full cycles where the script is tested on a fresh system to make sure it's still working. This is happening every time I do any sort of sysadmin-type of stuff. Even in ansible, chef, docker, ... To escape that something deeper needs to be changed.