About Searchplex ID

Searchplex ID is an open terminology registry for information retrieval and search engineering.

It assigns persistent identities to retrieval concepts and helps reconcile the vocabulary used across:

• IR research
• practitioner writing and talks
• industry engineering
• search-technology vendors
• search-system documentation

The goal is simple:

Make it easier for humans and machines to tell when different terms mean the same thing, when they are only closely related, and when they should remain distinct.

Why it exists

Retrieval terminology is fragmented.

Academic IR, production search teams, vendors, and documentation often use different names for the same or adjacent ideas.

Terms such as:

• query expansion
• query rewriting
• query reformulation
• candidate retrieval
• candidate generation
• first-stage retrieval
• first-stage ranking
• learned sparse retrieval
• semantic search
• late interaction
• multi-vector retrieval

are frequently used with overlapping or inconsistent meanings.

Searchplex ID provides stable concept identities so those vocabularies can be aligned without flattening meaningful distinctions.

What an ID represents

A Searchplex ID identifies a concept, not a webpage or a spelling.

For example:

https://id.searchplex.net/query-expansion/

identifies the concept Query Expansion.

The human-readable page is one representation of that concept. Machine-readable JSON and JSON-LD representations are also published.

Preferred names may evolve. The identifier remains stable.

What Searchplex ID publishes

Each concept may include:

• a preferred label
• alternative labels
• community-specific terminology
• a concise definition
• a scope note
• relationships to nearby concepts
• explicit non-equivalence where concepts are commonly conflated
• mappings to external vocabularies such as CSO, ACM CCS, or Wikidata
• a small number of defining or official links

Each concept may include mappings to established external vocabularies. Searchplex ID builds on standards such as SKOS and, where appropriate, maps retrieval concepts to authorities including CSO, ACM CCS, and Wikidata rather than duplicating them.

What it is not

Searchplex ID is not:

• an encyclopedia
• a paper database
• a leaderboard
• a claims or evidence graph
• a complete ontology of information retrieval
• a replacement for CSO, ACM CCS, Wikidata, or other established vocabularies

Its focus is narrower:

retrieval terminology and identity resolution.

Editorial approach

Searchplex ID is curated across several vocabulary environments:

• academic IR literature
• practitioner sources
• industry engineering publications
• search-technology vendor publications
• implementation documentation

These sources inform editorial decisions, but the registry publishes the normalized terminology rather than a full evidence trail.

When terminology is uncertain, concepts remain draft until the distinction is clear enough to publish.

Open data

Searchplex ID vocabulary data is published for open reuse under CC0.

Human-readable pages, plain JSON, JSON-LD, and bulk machine-readable exports are generated from the same canonical concept records.

Maintained by Searchplex

Searchplex ID is built and maintained by Searchplex, a retrieval engineering company.

The project is intended as a useful public resource for the search and information retrieval community.

Contributions and corrections are welcome through the project repository.