🚀 A Language Was Born Today: SMS Went Native

On March 24, 2026, SMS crossed a line.

What started as scripting glue is now a compiled native language path.

The Stack

• SML describes UI and scenes declaratively.
• SMS describes behavior and events.
• Runtime is native C++ under the hood.

The original goal was simplicity.
The side effect was performance.

What Changed Today

We implemented a full compiler path:

• SMS -> LLVM IR -> native code
• local trusted scripts run through AOT
• remote/untrusted content can stay on interpreter fallback
• CLI path works end-to-end (.sms -> executable)

And yes, we ran the Easter egg proof:

fun getAnswer() {
    return 42
}

fun main() {
    log.success("The answer of all questions is: ${getAnswer()}")
}

main()

Output:

Real Numbers (Same machine, same workload)

Protocol:

• 5 warmups
• 20 measured runs
• median + p95

Results:

• SMS Native: median 3466 us, p95 3502 us
• C#: median 4382 us, p95 4436 us
• GDScript: median 258220 us, p95 261709 us

Speedups:

• SMS Native vs C#: 1.26x (median), 1.27x (p95)
• SMS Native vs GDScript: 74.50x (median), 74.73x (p95)

That is not a tweet-sized benchmark claim.
That is measured, repeatable data from the same workload.

Why This Matters

I didn’t design a language around memory ceremony.

I did not specify:

• manual malloc/free in user scripts
• a tracing GC model for script authors

Yet we now have:

• simple script syntax
• native execution path
• reproducible benchmark gains
• fallback safety when needed

For me personally, this is huge:
my own computer language, 42 years after I first saw Amiga BASIC.

Built with love, coffee, and a stubborn focus on simplicity.

———

If you want, I can publish the benchmark harness and protocol in a follow-up post so others can
reproduce or challenge the numbers.