Updated documentation

README.md

@@ -1,17 +1,6 @@
 # ocean
 
-In-memory key value store that saves your data in JSON format.
-
-```
-1
-{"name":"User 1"}
-2
-{"name":"User 2"}
-3
-{"name":"User 3"}
-```
-
-If you like, you can operate on your entire data with classic UNIX tools.
+In-memory key value store that saves your data to disk using JSON.
 
 ## Installation
 
@@ -22,16 +11,16 @@ go get git.akyoto.dev/go/ocean
 ## Example
 
 ```go
-// User data
-type User struct {
-    Name string
-}
+// User type
+type User struct { Name string }
 
-// Load existing data from ~/.ocean/User.dat
-users := ocean.New[User]()
+// Create a new collection
+users := ocean.New[User]("todolist")
 
-// Store in memory and also store in ~/.ocean/User.dat
+// Store some data
 users.Set("1", &User{Name: "User 1"})
+users.Set("2", &User{Name: "User 2"})
+users.Set("3", &User{Name: "User 3"})
 
 // Read from memory
 first, err := users.Get("1")
@@ -42,17 +31,22 @@ for user := range users.All() {
 }
 ```
 
-In a real project you would usually prefix your collections with a project or company name:
+Data will be stored in `~/.ocean/todolist/User.dat`.
 
-```go
-// Data saved to ~/.ocean/google/User.dat
-users := ocean.New[User]("google")
+## Example file: User.dat
+
+```
+1
+{"name":"User 1"}
+2
+{"name":"User 2"}
+3
+{"name":"User 3"}
 ```
 
-Disk writes are async and they work like this:
+## Usage
 
-1. Set key and value in memory (sync.Map.Store)
-2. Mark the collection as "dirty" (atomic.StoreUint32)
-3. Immediately return control to the program
-
-Because a `Set` call doesn't immediately flush the memory to disk, calling `Set` multiple times in a web server request becomes extremely efficient.
+1. Create all the collections you need at the start
+2. `defer users.Sync()` to ensure queued writes will be handled on exit
+3. Start your web server
+4. Retrieve and update your data using `Get` and `Set` calls