Lexicon Convention Discussion

[lomky]

A ticket to discuss the conventions surrounding Lexicons.

Current Lexicon Conventions. (none)

Lexicon Fields:

gcis=# \d+ lexicon
                                                                 Table "gcis_metadata.lexicon"
   Column    |       Type        | Modifiers | Storage  | Stats target |                                      Description
-------------+-------------------+-----------+----------+--------------+---------------------------------------------------------------------------------------
 identifier  | character varying | not null  | extended |              | A brief descriptive identifier for this lexicon.
 description | character varying |           | extended |              | A description of the lexicon, possibly including the organization associated with it.
 url         | character varying |           | extended |              | A url for further information.
Indexes:
    "lexicon_pkey" PRIMARY KEY, btree (identifier)
Check constraints:
    "lexicon_identifier_check" CHECK (identifier::text ~ similar_escape('[a-z0-9_-]+'::text, NULL::text))
Referenced by:
    TABLE "exterm" CONSTRAINT "exterm_lexicon_identifier_fkey" FOREIGN KEY (lexicon_identifier) REFERENCES lexicon(identifier) ON UPDATE CASCADE ON DELETE CASCADE
Triggers:
    audit_trigger_row AFTER INSERT OR DELETE OR UPDATE ON lexicon FOR EACH ROW EXECUTE PROCEDURE audit.if_modified_func('true')
    audit_trigger_stm AFTER TRUNCATE ON lexicon FOR EACH STATEMENT EXECUTE PROCEDURE audit.if_modified_func('true')

Provenance Connections:

none

Relationships:

none

[lomky]

Lexicon Conventions

Current Use Case

In the past, we used these for collecting equivalent acronyms from outside systems.

We don't actively seek out new lexicons, but are open to mapping in new ones.

Issues & Thoughts

We don't currently note the version of the lexicon.
How do we keep these lexicons up to date? Not let them go dead?
We aren't sure how important these are to GCIS in an ongoing manner.
Acronyms are a good possible use case.
We should do a bit of digging into the ones we have to determine which are worth maintaining up to date.

Datagov Convention

• Make sure to grab both the opaque identifier and the human-readable identifiers when creating new datasets.