Astrolabe定盤星

Educational Research and Technology Centre, The Hong Kong Polytechnic University

Astrolabe定盤星

An instrument for locating your teaching

An astrolabe shows the structure of the sky and tells you where you stand in it. It does not tell you where to go, and neither does this.

Astrolabe teaches educators to understand, distinguish, select among, and practise every form of “X-based” education.

About a minute. There is no account yet, and nothing you click here can be joined to one later.

The distinction the system turns on

Two kinds of operation, and every screen says which

Every operation Astrolabe performs is registered as one of two things. Anchored operations have no path to a model at all — not a refusal at the surface, but an absence in the wiring, which a test asserts.

Extendable

A model may do this

Reading a document to spec, posing a contrast case, generating a discrimination item, critiquing something you wrote.

Anchored

A person must do this

Your reasoning, your design attempts, your judgements of quality, and your decisions about what your students must do themselves.

Before you sign up

Three questions about teaching, and what the evidence actually says

Each one is an authored finding with its sources. Where the evidence is contested you get both cases rather than a verdict.

One · empirical

Instructors were asked to predict how their own students would do on a concept question before the students answered it.

How close were the predictions?

Two · conceptual

Competency-based, project-based, mastery, inquiry-based, work-based, flipped, team-based. Seven labels people use as though they were alternatives to one another.

How many of these can be true of one course at the same time?

Three · urgent

Take the main written assessment in your course. Suppose a model produces a submission that meets every criterion on your rubric.

What follows?

About a minute. There is no account yet, and nothing you click here can be joined to one later.

Stated limits

What Astrolabe does not do

Each of these is a test in the suite rather than a paragraph in a policy. The name beside it is where it is enforced.

1

It does not rank pedagogies. Layer 3 eliminates on constraints and orders survivors by how near they are to what you already do.

Enforced in eliminationService
2

It does not draft your work. No field where you design offers a generated draft, and there is no setting that adds one.

Enforced in designService.assertNoGenerationAffordance
3

It does not issue a verdict where the evidence is contested. You get both cases and the question that would settle it.

Enforced in claims.renderableClaim
4

It does not certify teaching competence. It supplies evidence; an institution that wants a credential warrants that itself.

Enforced in reportService.buildEvidenceRecord
5

It holds no hidden profile. Every inference we hold is readable, correctable and exportable by you.

Enforced in /app/evidence · /app/export
Three ways in

Who it is for

None of these roles is a superset of another. A training head sees things an administrator cannot, and the reverse.

Learner

Redesign one course

You are redesigning one course, and you want to know what you are actually doing before you change it.

  • A profile built from your own materials and your own reasoning
  • Approaches eliminated on your constraints, survivors ordered by nearness
  • Every inference readable, correctable and exportable
Sign in →
Training head

Run the cohort

You run the programme, and you need to know who to write to this morning.

  • A cohort sorted by what you would have to do
  • Anyone left mid-air, named as the portal’s failure
  • Cross-disciplinary trios, formed when people are ready
Sign in →
Administrator

Keep the model honest

You keep the pedagogical model honest, and you answer for what it asserts.

  • The claims register, with strength, sources and disputes
  • Per-descriptor agreement, and what may not be reported alone
  • Provider routing, consent and the ethics reference
Sign in →
Computed as this page renders, so it cannot drift

The instrument reports its own state

6
Genuine rivalries in the taxonomy — pairs that cannot both be true of one decision
18
Descriptors, each with its own agreement evidence
20
Pedagogical claims on the register, disputes included
PROVISIONAL — synthetic figures pending the panel study. Public launch is gated on the real study (§1b.2), not on this file.
The state of the agreement study behind the coding scheme
Language
Appearance