Curriculum In use

Curriculum & Standards

bina composes its own curriculum from the national curricula of four countries and the international frameworks alongside them, and holds the result as data: every standard is a node in a graph carrying the source items it was drawn from, the sub-skills it decomposes into, and the written rules that placed it where it sits. Publication checks and teacher review govern what reaches the classroom. bina owns its curriculum the way a software company owns its code.

Those rules are themselves compiled. The published record of learning science (papers, research reviews, institutional guides, textbooks) is admitted into a corpus, broken into individual claims with their evidence and confidence attached, and compiled into the ruleset the pipeline answers to, with a named owner accepting each revision. Research about how children learn can arrive in a classroom here as a reviewed rule that can be walked back to the page it came from.

The shape of the curriculum

This representative view shows how the standards connect. Distance from the core is the Level, so Foundation sits at the center and middle school reaches the surface, and each of the seven disciplines grows outward through its own region. All five relation types the model carries are drawn. Drag to turn it. Hover a node to light up the development path it belongs to.

Distance from the center is the Level: Foundation at the core, middle school at the surface, with each of the seven disciplines growing outward through its own region. Five relations are drawn between standards: progression along a strand, prerequisite across strands of one discipline, cross-discipline links cutting through the sphere, equivalence, and decomposition into sub-skills.

A representative slice of the graph, laid out and rendered live in the browser. The structure is bina’s: distance from the core is the Level, running Foundation to middle school; the seven disciplines each grow outward through their own region; and all five relation types the model carries are drawn: progression, prerequisite, cross-discipline, equivalence and decomposition. One development path at a time is traced outward from its earliest Level, which is how a strand actually accumulates. Strand names here are the shared vocabulary of the national frameworks; bina’s own standard text and codes stay off this page.

What the engine holds

One standards graph across seven disciplines and every Level from Foundation through middle school, composed under a versioned ruleset and queryable in both directions.

The standards graph bina’s own standards as first-class nodes, joined by five typed edges: prerequisite, progression, decomposition, equivalence and cross-discipline.

Depth in the graph emerges from prerequisite chains, so the structure carries the hierarchy with no fixed layers imposed on top. The same model holds a largely linear discipline like mathematics and a web-shaped one like literacy, with no special-casing between them.

Framework anchors Every published standard maps to external framework standards many-to-many, each link carrying a match type and a confidence.

External standards are reference data anchored to bina’s own, so four national frameworks all resolve against a single graph. The lookup also runs in reverse: given any standard in a national framework, the engine returns the bina standards that cover it.

Micro-credentials Each standard decomposes into measurable sub-skills through a documented task analysis, within bounded grain size.

The standard is the umbrella that carries scope, transferability and social-emotional integration; the sub-skill is the operational unit that carries the pedagogical detail and is what a child is actually assessed against. Decompositions that come out too coarse or too fine are flagged for review.

The research corpus Stored sources, one processing record each, atomic insights carrying their own evidence and confidence, and thematic indexes across the whole body.

Research papers, institutional guides, textbooks and framework analyses are held verbatim and never edited. Each is processed into a record naming who read it and what was extracted, then broken into single claims tagged with the disciplines, the Levels and the pipeline stages they bear on. Any claim in the curriculum walks back to the document it came from, and any document follows forward to the rules it changed.

Encoded pedagogy The Adventure Approach, the developmental profile for each Level, the social-emotional model and the technology boundaries, authored as documents the pipeline reads.

Unit shape, lesson rhythm, the non-negotiables for every learning adventure, and each institutional value translated into a concrete design requirement. Their rationale is recorded and reviewed alongside external evidence, without treating an institutional principle as a research finding.

The policy ruleset Three tiers of written rule: hard rules that block publication, soft rules that score quality, and innovation rules that push the curriculum forward.

Each rule names the stage that evaluates it, the entity it applies to, the condition that triggers it and the rationale behind it. The ruleset is versioned, owned by the curriculum lead, and recompiled in a recorded pass whenever the corpus behind it moves.

Playbooks Operational decision procedures for the judgment calls that recur: grade placement, merge strategy, sub-skill design, per-discipline sequencing.

Each playbook carries a confidence level and a decision trail, and grows every time a live composition run produces a call worth keeping. Curriculum expertise ends up in the repository as a runnable procedure.

Sequencing and pacing Standards grouped into eight cohesive teaching periods a year per discipline, four to five weeks each, balanced against the weekly hour budget for the Level.

Placement is solved against the teaching hours a Level actually has, and against the prerequisite edges already in the graph, so a period cannot ask for a skill the child has not been taught. Four to five weeks per unit is a pedagogical invariant: enough time for depth, then consolidation.

The navigator The whole curriculum as a browsable three-dimensional graph, with search and filtering across thousands of standards.

Curriculum designers and teachers read the graph as a tech tree: what unlocks what, where a discipline crosses another, how a strand progresses through the Levels. Full-text and fuzzy search run in the store itself, so a designer finds a standard by half-remembering its wording.

Three ways a school gets a curriculum

Schools can adopt a national framework, author a scheme of work in documents, or build a connected model of their own. bina composes: the unit of curriculum here is a node in a graph with provenance attached, and that is what makes everything downstream of it addressable.

	Adopted whole	Authored in documents	bina: the composed graph
Unit of curriculum	A national framework, taken as published.	A scheme of work per subject.	A standard: a graph node carrying its source items, its sub-skills, its edges and the rules that placed it.
Where it lives	With the framework’s owner.	In documents and spreadsheets.	In a database the school owns, versioned, and queryable in both directions.
How a standard is justified	By the framework’s authority.	By the author’s professional judgment and documented sources.	By provenance. A standard publishes only while it carries at least one external source item and a recorded reason for every decision that moved it.
Cross-framework equivalence	One framework, one answer.	Mapped and maintained separately.	Many-to-many anchors with match type and confidence on every link, resolvable from either end.
Changing it	Wait for the owner’s revision cycle.	Edit the document, then review the downstream effects.	Recompose the affected slice. The hard gates re-run, and the write lands atomically or leaves the live graph untouched.
What new research changes	Updates depend on the framework and the school’s own review.	Research informs the author’s revision process.	A source enters the corpus, becomes insights, compiles into rules, and the next composition run inherits them across every discipline at once.
What a family can be told	“We follow one country’s curriculum.”	“We follow our own scheme of work.”	“Here is your child’s position against the framework you care about”, answered from the same graph the lesson was built from.
The trade	Someone else’s sequencing, someone else’s assumptions about how children develop.	Quick to start; coherence needs deliberate review as it grows.	The cost is time. A discipline enters the graph only once it has been composed, reviewed by teachers and passed the gates, so the curriculum grows one discipline at a time and bina accepts that pace for what it buys.

From framework document to published standard

Every stage consumes the previous stage’s artifact and commits its own, and the policy ruleset is evaluated at every one of them, with a final suite standing between composition and the classroom. The corpus closes the circuit: teacher review, new research and pipeline overrides compile back into the rules that govern the next run.

Policy rulesetHover or tap any checkpoint on the rail to see what it enforces.

Ruleset checkpoints. narrow: the rules decide what may be compared at all. adjudicate: pairwise alignment re-run with the order swapped, every decision citing its rules. compose: hard rules applied while standards, sub-skills and edges are built. balance: placements checked against actual weekly hours and progression order. verify: the publish suite (acyclic graph, provenance, hard rules, placement consistency, load, cluster stability, referential integrity). publish: a scoped atomic write, or the live graph is left untouched.

Composition is a governed run with written policy behind every decision in it. The policy ruleset is evaluated at four points inside the pipeline and again as a publish suite, so questionable placements can be caught before publication. What teachers say in review, what new research establishes, and what the pipeline itself had to override all return to the corpus, which is recompiled into the ruleset the next run answers to.

How the frontier of research becomes a rule

The return path on the diagram above opens up like this. The published record of learning science (papers, research reviews, institutional guides, textbooks, standards references) is admitted into a corpus, broken into individual claims with their evidence attached, cross-indexed, compiled into rules, and accepted by a named owner before the pipeline is allowed to answer to it. Research reaches a classroom here through one door, and the door keeps a record.

The corpusHover or tap any step to see what it produces.

Process: sources stored as published and never edited, one processing record each naming what was taken and what was left; sources confirmed but still unread sit in the same index with an owner against each. Extract: one claim to a file, with an evidence table, a confidence grade and the routing tags (disciplines, Levels, and the pipeline stages that should see it), ending in a written implication for the composition procedure. Index: insights cross-filed under themes, so a theme accumulates across every source that touched it. Compile: a recorded pass audits the backlog and drafts a ruleset revision, every rule citing the insights behind it, each section carrying its own last-compiled record. Accept: the curriculum lead reads the proposal and accepts it, and only then does the pipeline answer to it.

The corpus is a separate body from the national frameworks the pipeline composes. Frameworks supply the material a standard is made of; the corpus supplies the rules that decide how it is made. bina’s own pedagogy (the Adventure Approach, the developmental profile for each Level, the social-emotional model, the technology boundaries) is admitted through the same processing as an external paper and held to the same evidence discipline, which is why an institutional value arrives at the pipeline as a rule with a rationale attached.

How a standard earns its place

Four properties carry most of the weight here: provenance, a written ruleset, order-independent merging, and a review loop that ends in the corpus.

Provenance is a condition of publication

A standard publishes only while it carries at least one external source item it was drawn from. The check is a hard gate, so a plausible-sounding standard with nothing behind it cannot enter the graph. Any placement that departs from the default carries a written reason beside it.

bina’s own pedagogy is processed like a paper

The Adventure Approach, the developmental profile for each Level, the social-emotional model and the technology boundaries enter the corpus as sources, are read into single claims, graded for confidence, and compiled into the same ruleset external research compiles into. An institutional value arrives at the pipeline as a rule with a rationale, which is what makes it possible to argue with.

Merge decisions survive their own order

Frameworks are aligned in pairs, then the whole comparison re-runs with the pairing swapped. Clusters that come out the same both ways are accepted; the ones that move are flagged for a person. Artifacts of processing order stay out of the curriculum.

Transferability is the default placement rule

A standard sits at the earliest age any source framework introduces it, to support continuity across frameworks; an individual child’s starting point still needs assessment. Holding one back requires a prerequisite constraint or a recorded rationale, and the rationale stays attached.

The same concept at two depths stays two standards

A spiral curriculum revisits a concept at rising complexity on purpose. Two standards collapse into one only when the concept and the complexity level both match, which keeps deliberate repetition from being optimized away as duplication.

Teacher review ends up in the corpus

Comments from the teaching team on composed standards are synthesized with the reasoning behind them and entered as corpus records. The next recomposition inherits them, so a judgment can inform later disciplines where the same reasoning applies.

What compounds

One composition, four national answers

Because anchors are many-to-many and resolve from either end, a single graph answers “where is my child against the UK?” and the same question against the US, Canada and Australia. Adding a framework means anchoring it onto standards that already exist, so the curriculum stays one graph as its reach grows.

The graph is the substrate for everything downstream

Sequencing, unit composition, coverage and gap analysis all read the same nodes and edges. Each capability built on top inherits the provenance, the anchors and the sub-skill decomposition already there, so it opens against a curriculum that is machine-readable from the first day.

New evidence has a route into practice

A source can be processed, its insights proposed as rules, and a reviewed revision carried into the next composition run. Updating a connected curriculum still takes judgment; the useful difference is that the route from evidence to change is explicit.

Built on A relational store holding the graph, with recursive traversal, full-text and fuzzy search running inside the database itself. A Go service exposes the graph as a JSON API and as an MCP tool surface, so a curriculum designer and an agent operate it through the same validated handlers, and every mutation is a typed operation with an audit trail. A browsable three-dimensional view sits on top. The corpus, the rulesets and the playbooks are plain versioned documents in the same repository as the code that reads them.

Where this goes next

Every standard already carries the sub-skills it decomposes into and the frameworks it anchors to, which is exactly what a progress record needs to attach to. The same graph makes a child’s position expressible against bina’s own standards and answerable against whichever national framework a family cares about, drawn from the same data the lesson was composed from, with the evidence trail intact.