Refactoring English: A mathematically compressed dialect to bridge the gap between human and machine logic.

As developers, we spend all day writing strict, strongly-typed logic. If an API returns a vague string instead of a boolean or a defined integer, the system breaks.

Yet, when we step away from the IDE and talk to each other—or increasingly, when we prompt Artificial Intelligence—we use a legacy system riddled with bugs: Standard English.

We use "null" values like "soon." We use subjective variables like "probably." We use floating, scaleless metrics like "almost done." AI has exposed just how inefficient our language is. When you prompt an LLM with subjective idioms, it hallucinates timelines. When humans speak to each other with them, we misalign on expectations. To bridge the gap between human thought and machine execution, we need an API for human communication.

So, I built one.

Englicode is a mathematically compressed, Base-10 dialect of English. It treats daily communication the way we treat software inputs: strict, measurable, and scalable.

Here is how the architecture works, using three core protocols.

1. The Bandwidth Protocol (Replacing Static Percentages)

Percentages are flawed because they lack scale. "50% capacity" means nothing if you don't know the ceiling. Are we talking about 10 users or 10,000 users?

Englicode replaces percentages with a Dynamic Anchor syntax: [Current Value] [Anchor].

Standard English: "We've used half the server RAM."

Englicode: "Server RAM is 20 40." (Logic: It instantly communicates the 50% load, but more importantly, it proves we still have 20 units of headroom).

If a system breaches capacity, the syntax naturally scales to show the exact overflow:

Englicode: "Budget is 60 50." (Logic: Critical Overload. We exceeded the 50-unit anchor by exactly 10).

1. The Time Index (Deprecating "Soon")

"I'll do it soon" is the worst string in the English language. It means 5 minutes to a manager, but 5 hours to a developer. Furthermore, the 60-minute/12-hour clock forces our brains into unnecessary fraction conversions.

Englicode maps time to a clean, 1-to-7 Base-10 Index. (1=Seconds, 2=Minutes, 3=Hours, 4=Days, 5=Weeks).

Syntax: [Value] [Time Index]

"I'll deploy the fix in a couple of hours." → "Deploying in 2 3." (2 Hours).

"I hit the gym three times a week." → "Gym frequency is 3 5." (3 occurrences per Week).

It replaces paragraphs of scheduling text with two integers.

1. The Certainty Protocol (Boolean Probability)

In software, truth is binary (0 or 1). In human probability, we use messy words like "maybe" or "highly likely."

Englicode forces Boolean-adjacent thinking to quantify risk. 0 is Impossible. 1 is Absolute Truth.

Syntax: [Decimal] 1

"No way / Impossible" → 0 1

"Maybe / 50-50" → 0.5 1

"Definitely / Confirmed" → 1 1

Instead of telling your team, "The database migration will probably work," saying "Migration certainty is 0.8 1" forces you to quantify your confidence and explicitly admit a 20% chance of failure.
The AI Payload: Zero-Shot Translation

If a logic system is truly universal, machines should understand it instantly. Because Englicode strips away cultural bloat and relies entirely on integers and arrays, LLMs parse it perfectly without custom training.

I built a machine-readable schema for the language. You can simply prompt ChatGPT or Claude with:

"Read the syntax rules at englicode.com/schema.json and apply them to this conversation."

Once the AI ingests the schema, you can prompt it in raw math: "Draft an update to the client. Project status is 80 100. Delivery is in 2 4." The AI will instantly translate your mental shorthand into a perfectly formatted, polite corporate email ("The project is 80% complete and will be delivered in 2 days").
It's Open Source (Submit a PR)

Englicode isn't a closed SaaS tool; it's an open-source logic experiment.

The dictionary is governed by a strict "Pull Request" system via a Consensus Board. If you think the Base-10 Time Index is mathematically flawed, or if you have a better protocol for spatial distance, submit a PR to the repository and prove the math. If the community upvotes the logic, it becomes canon.

I've also built a "Cognitive Gym" on the site—a series of rapid-fire translation quizzes that test your processing speed against these new metric protocols.

You can read the full documentation and test your logic here:

The Web Sandbox & Quizzes: englicode.com

The AI Schema: englicode.com/schema.json

Comprehensive Docs: github.com/tngoman/englicode

I know speaking in strict data types is jarring at first. But once your brain gets used to compressing reality into pure logic, standard English just feels like writing in Assembly.

Give the sandbox a try, and let me know where my logic fails.