ref: rewrite Typical Workflow in index page

rel. #3223 (comment)

content/docs/command-reference/index.md

@@ -10,16 +10,14 @@ does not change directories in your terminal).
 
 ## Typical DVC workflow
 
-- In an existing Git repository, initialize a <abbr>DVC project</abbr> with
-  `dvc init`.
-- Copy data files or dataset directories for modeling into the repository, and
-  track them with DVC using the `dvc add` command.
-- Process the data with your own source code, using `dvc.yaml` and/or the
-  `dvc stage add` command to specify further <abbr>outputs</abbr> that should
-  also be tracked by DVC, and executing the code using `dvc repro`.
-- Sharing a <abbr>DVC repository</abbr> with the codified data
-  [pipeline](/doc/command-reference/dag) will not include the project's
-  <abbr>cache</abbr>. Use [remote storage](/doc/command-reference/remote) and
-  `dvc push` to share this cache (data tracked by DVC).
-- Use `dvc repro` to automatically reproduce your full pipeline iteratively as
-  input data or source code change.
+- Initialize a <abbr>DVC project</abbr> in a Git repo with `dvc init`.
+- Copy data files or dataset directories for modeling into the project and use
+  `dvc add` to tell DVC to <abbr>cache</abbr> and track them.
+- Create a simple `dvc.yaml` file to codify a data processing
+  [pipeline](/doc/command-reference/dag). It uses your own source code and
+  specifies further data <abbr>outputs</abbr> for DVC to control.
+- Execute or restore any version of your pipeline using `dvc repro`, or
+  experiment on it with `dvc exp` features.
+- Sharing the <abbr>repository</abbr> will not include locally cached data. Use
+  [remote storage](/doc/command-reference/remote) with `dvc push` and `dvc pull`
+  to share data artifacts.