Create 0008-io-operations-naming-conversions.md

docs/0008-io-operations-naming-conversions.md

@@ -0,0 +1,53 @@
+# I/O operations naming conventions
+
+* Status: Accepted
+* Deciders: Marino, Zorica, Stefanija, Davide, Tomasz, Maciej
+* Date: 22/09/2021
+
+
+## Context and Problem Statement
+
+We would like to have well defined conventions about naming for I/O functions ( mainly DAOs, APIs, Repositories, Use Cases ), for the sake of consistency.
+
+## Considered Options
+
+For **DAOs**:
+
+* `get-`
+* `find-`
+* `fetch-`
+
+For **APIs**
+
+* `get-`
+* `fetch-`
+
+For higher levels ( **repositories, use cases, workers** )
+
+* `get-` / `get~Once`
+* `observe-`
+
+## Decision Outcome
+
+Every component will respect the global conventions of the underlying language ( SQL for database and HTTP for network )
+
+For **DAOs**
+
+* `find-` for single item
+* `findAll-` for many items
+* `insert-` / `update-` / `set-` for write
+
+For **APIs**
+
+* `get-` for single and many items
+* `put-` / `post-` 
+
+For higher levels ( **repositories, use cases, workers** )
+
+* `get-` to *get* a single event ( see suspend function )
+* `observe-` to *observe* a stream of events ( see Coroutines Flow )
+* Common sense general developing names for actions / write ( like `moveMessages`, `markAsRead`, etc )
+
+## Consequences
+1. All the new I/O functions declared in the application should follow the conventions defined above
+2. Whenever possible, we should rename the pre-existen ones