T
The topology of SKOS is set-theoretic rather than strictly hierarchical, which is much less restrictive and more enlightened take on organizing information.
Other
Port SKOS (RDF) to ATProto Lexicon.
ATProto is a thing that happens over time. Are users actually specifically looking for ATProto-enabled apps? As of late January 2026, there are over 1000 lexicons listed in lexicon.garden, and that number is steadily rising. Create a lexicon for concept schemes in general. Create a taxonomy for ATProto app themes. Create an index of ATProto lexicons by theme and/or semantic content. Crib from the RDF ecosystem, which has more or less figured this out. Determine the extent to which Lexicon and RDF are isomorphic, or at least that the expressive range of Lexicon is a subset of the range of RDF. Developers benefit by increased findability. Developers new to the ecosystem may not know that a lexicon that suits their needs already exists. Even if an existing lexicon provided partial coverage, a developer could extend it via subclass or mix-in. For one, you could publish the taxonomy and the app index to the protocol. How do we represent the themes? How would we represent the taxonomy itself? If SKOS is portable to Lexicon, could other RDF vocabularies also be portable? Is there prior art on indexing data vocabularies by theme? Is there prior art we can use for a concept scheme lexicon? Lexicons represent a larger common factor than individual apps (if e.g. an app uses a health lexicon, it's probably a health app). Model the changes to index membership as well as changes to the topology of the taxonomy itself. Port SKOS (RDF) to ATProto Lexicon. SKOS concepts are identifiable, addressable data objects which are distinct from their labels. SKOS is a stable ontology and its proponents have already fairly exhaustively plumbed the problem space. SKOS is already deployed in e.g. content management systems. Semantic collisions are already probably occurring among existing lexicons, and will likely increase as more are developed. Tag lexicons, rather than apps, with taxonomy concepts. Taxonomy concepts would exist as addressable objects within ATProto. The topology of SKOS is set-theoretic rather than strictly hierarchical, which is much less restrictive and more enlightened take on organizing information. Users are coming to the ATmosphere looking for ATProto-compatible apps in various application domains. Users can browse ATProto apps from familiar categories, without being overwhelemed by choices. Why an index of lexicons and not an index of apps?