DASH vs Tcl

Dash has a very fast startup, this happens because the shell is started a lot of times during boot and dash minimizes the work it does during this process.

It is designed to be very lightweight and has no support for shell specific extensions that are not POSIX.

DASH is the default shell for Debian based systems due to it speed, full POSIX compliance and low overhead.

It's fully POSIX compatible, so if your script runs on dash it will probably run on all other shells.

It's a clone of the original System V4 Bourne shell.

Unlike *sh Tcl has a rich standard library.

Unlike in sh you don't need to quote your variable substitutions ('"$1"').
Expanding of arguments occurs mostly explicit and if yet expected in-place (e. g. using eval or subst) it follows strict and well clear rules.
So exec test $argv would execute test with single argument.
And appending {*} before $argv would execute test with all arguments (list expansion).

With a little care you can have the same script work on Linux, *BSD, OS X and Windows.

You can expect a reasonably recent version of Tcl to either installed or available in the repositories of any popular open source *nix.

Tcl can operate at the same level of abstraction as the POSIX shell, which makes it easier to manipulate the output of other programs.

Tcl enables easy deployment through self-contained binaries known as starpacks.

Want to run something 5 times? Here you go: set i 0; time { puts done-[incr i] } 5
If you need real conditional cycle? Not a problem: for {set ready 0; set i 1} {$i <= 100 && !$ready} {incr i} { if {[exec do-some-thing] eq "ready" } {set ready 1} }
Want to measure performance of something or repeat it max 300 times and not longer than 1 seconds? Very simple: timerate { after 20 } 1000 300
How about notifying yourself when some http-server is back online? Sure thing: while {[catch { close [socket localhost 80] }]} { after 1000 }; puts "\7\7\7ONLINE!"
And you can do it also fully asynchronously using events etc.