Improve README

Remove typos, improve sentences

README.md

@@ -7,9 +7,9 @@
 > One kubectl to bring them all  
 > and in the darkness bind them.  
 
-Managing different kubernetes clusters ofter requires to keep multiple versions
-of the kubectl available on the system, plus it requires some way to ensure the
-right binary is used when talking with a cluster.
+Managing different kubernetes clusters often requires to keep multiple versions
+of the kubectl available on the system, plus it poses the challenge to ensure
+the right binary is used when talking with a cluster.
 
 kubernetes defines a clear [version skew policy](https://kubernetes.io/docs/setup/release/version-skew-policy/)
 for all its components. This is [what is stated about kubectl](https://kubernetes.io/docs/setup/release/version-skew-policy/#kubectl):
@@ -23,13 +23,16 @@ for all its components. This is [what is stated about kubectl](https://kubernete
 > kubectl is supported at 1.19, 1.18, and 1.17
 > ```
 
-kuberlr (*kube-ruler*) is a simple wrapper for kubectl. It's main purpose is to
+kuberlr (*kube-ruler*) is a simple wrapper for kubectl. Its main purpose is to
 make it easy to manage clusters running different versions of kubernetes.
 
-This is how kuberlr looks like:
+This is how kuberlr looks like in action:
 [![asciicast](https://asciinema.org/a/326626.svg)](https://asciinema.org/a/326626)
 
-## Usage
+## Installation
+
+You can find pre-built binaries of kuberlr under the
+[GitHub release](https://github.com/flavio/kuberlr/releases) tab.
 
 Put the `kuberlr` binary somewhere in your `PATH` and create a symlink named `kubectl`
 pointing to it.
@@ -41,8 +44,10 @@ $ cp kuberlr /bin/
 $ ln -s ~/bin/kuberlr ~/bin/kubectl
 ```
 
-Now you can use `kubectl` as you would, behind the scene `kuberlr` will ensure
-the right version of `kubectl` is used.
+## Usage
+
+Use the `kubectl` *"fake binary"* as you usually do. Behind the scene
+kuberlr will ensure a compatible version of `kubectl` is used.
 
 ## How it works
 
@@ -54,25 +59,14 @@ kuberlr obtains the url of the kubernetes cluster either by looking at the
 the `KUBECONFIG` environment variable.
 
 Once the version of the remote server is know, kuberlr looks for a compatible
-version under the `~/.kuberlr/<GOOS>-<GOARCH>/` directory.
+kubectl binary under the `~/.kuberlr/<GOOS>-<GOARCH>/` directory.
 
 kuberlr reuses an already existing binary if it respects the kubectl
 version skew policy, otherwise it downloads the right one from the
 [upstream mirror](https://kubernetes.io/docs/tasks/tools/install-kubectl/).
-kuberlr downloads the kubectl binary to the following destination:
-`~/.kuberlr/<GOOS>-<GOARCH>/kubectl-<remote server major version>.<remote server minor version>.<remote server patch level>`
+
+kuberlr names the kubectl binaries it downloads using the following naming
+scheme: `kubectl-<major version>.<minor version>.<patch level>`.
 
 Finally kuberlr performs an [execve(2)](https://www.unix.com/man-page/bsd/2/EXECVE/)
 syscall and leaves the control to the kubectl binary.
-
-## TODO
-
-* [ ] I've tested kuberlr only on Linux. It should work also on macOS and on Windows.
-  Feedback is welcome.
-* [x] Provide pre-built binaries for kuberlr (WIP)
-* [x] Test coverage, code linting,... right now this is a toy project I created
-  during a weekend afternoon
-* [x] Relax the versioning constraint when downloading the kubectl version.
-  Right now accessing a kubernetes 1.16.0 and a 1.16.3 cluster would result in two
-  kubectl binaries being downloaded, while it would just be fine to use the latest
-  one