DPCH04 — Discoverable
“Registered in Data Marketplace or Catalog”
What DPCH04 is really asserting
DPCH04 is not asserting that:
“There is some metadata somewhere.”
It is asserting that:
A Data Product is intentionally published into an enterprise discovery surface where potential consumers can find, understand, and evaluate it without prior knowledge or personal connections.
Discoverability is about intentional exposure, not accidental visibility.
The Essence (HDIP + Data Mesh Interpretation)
A Data Product is discoverable if and only if:
- It is registered in an enterprise discovery mechanism (catalog and/or marketplace)
- It can be found by non-creators
- Its existence and purpose are decoupled from tribal knowledge
If consumers need to:
- ask around,
- know the team,
- or already know the product exists,
then DPCH04 is not met, regardless of how rich the metadata is.
Positive Criteria — When DPCH04 is met
DPCH04 is met when all of the following are true:
1. Product is explicitly published
The Data Product:
- appears in a catalog or marketplace
- has a stable identity and listing
- is visible to its intended audience (even if access is gated)
Publication is an intentional act, not a side effect of deployment.
2. Discovery works for unknown consumers
A consumer who:
- does not know the producing team
- does not know the system of origin
can still:
- find the product via search or browse
- understand what it is for
- assess whether it is relevant
This is the acid test.
3. Discoverability is product-level, not dataset-level
The unit of discovery is the Data Product, not:
- individual tables
- raw files
- pipeline outputs
The product may expose multiple outputs, but discovery happens at the product abstraction.
Negative Criteria — When DPCH04 is not met
DPCH04 is not met if any of the following are true:
❌ Product is “discoverable” only by insiders
Examples:
- shared Confluence page
- internal wiki
- Slack messages
- team-owned documentation
This is tribal discovery, not product discovery.
❌ Only technical artifacts are registered
Examples:
- tables appear in a metastore
- schemas visible in query engines
- files exist in object storage
These are assets, not discoverable products.
❌ Registration is automatic and opaque
Examples:
- pipelines auto-register tables
- metadata exists but product intent is missing
- no explicit “publish” step
If the domain has not chosen to publish, discoverability is accidental.
Edge Cases (Important Guidance for Agents)
Case 1: “Catalog entry exists but is meaningless”
⚠️ Partial
Rationale:
- Product exists in discovery surface
- But consumers cannot evaluate relevance
- Often transitional state
Case 2: “Marketplace listing with gated access”
✅ Met
Rationale:
- Discoverability ≠ accessibility
- It’s acceptable to see that a product exists even if access is restricted
Case 3: “Only one known consumer”
❌ Not met
Rationale:
- A single pre-arranged consumer does not imply discoverability
- The product must be findable by others
Evidence Signals an Agent Should Look For
Authoritative evidence:
- Catalog or marketplace listing
- Search index entry
- Product-level page with description and owner
Supporting evidence:
- Tags, domain classification
- Linked glossary terms
- Usage instructions
Red flags:
- Discovery relies on system names
- Listings auto-generated from pipelines
- No explicit “publish” action by DPRO
How an AI Agent Should Decide
Decision rule (simplified):
If a potential consumer cannot find and understand the Data Product without knowing the producing team, DPCH04 is not met.
Why DPCH04 Is Non-Negotiable
Without DPCH04:
- reuse cannot scale
- governance cannot see the product
- platform value remains invisible
- Data Mesh degenerates into distributed silos
Discoverability is the bridge between autonomy and reuse.
Canonical Statement (for BPS)
DPCH04 is satisfied only when a Data Product is intentionally published into an enterprise discovery surface where unknown consumers can find, understand, and evaluate it without prior relationships or tribal knowledge.