When to use $(...) instead of backticks?

That's a great question about a common point of confusion in shell scripting! You're asking about $() versus backticks (`) for command substitution.

Both $(command) and `command` perform command substitution, meaning they execute a command and replace themselves with that command's standard output.

However, $(command) is generally the preferred and more modern syntax. Here's why and when to use one over the other:

Why $(...) is preferred:

1. Readability: $(...) is much easier to read, especially when dealing with nested commands.

   • Good: tar -cvf backup-$(date +%Y%m%d).tar $(find . -name "*.txt")
   • Harder to read: tar -cvf backup-`date +%Y%m%d`.tar `find . -name "*.txt"`
     The backticks can get lost among other quotes or characters, making it harder to discern where a command substitution begins and ends.

2. Nesting: $(...) allows for easy and natural nesting.

   • Example with nesting: output=$(cat $(find . -name "file.txt"))
   • Nesting with backticks is tricky and requires escaping: output=`cat \`find . -name "file.txt"\ `` (Notice the \ before the inner backticks). This quickly becomes cumbersome and error-prone.

3. Clarity with other expanding characters: Backticks can sometimes be confused with single quotes. $(...) makes it clear you're performing command substitution.

When you might still see or use backticks:

• Legacy Scripts: Many older shell scripts use backticks because $(...) was introduced later (though it's been standard in POSIX shells for a long time, including Bash). You'll encounter them when working with existing codebases.
• Simplicity: For very simple, non-nested commands, some people might still use backticks out of habit.

In summary:

• Always use $(...) for new scripts or when modifying existing ones. It's more readable, easier to nest, and less prone to errors.
• Be aware of backticks (`) because you will see them in older scripts, but aim to use $(...) yourself.

In your current step, POSITION=$(expr index "$STRING" "$CHAR") uses the recommended $(...) syntax for good reason – it clearly captures the output of the expr command.

Does that explanation clarify the difference for you?