Readme update

README.md

@@ -24,15 +24,34 @@ rolling dependency for continuous integration.
 get nix through your package manager or `curl https://nixos.org/nix/install |
 sh`
 
-## usage
+## usage (reccomended)
 
-talk to @vreis, or  
+talk to @vreis, or:
 
-use `argopkgs` as a channel:
+```
+$ git clone git@xgitlab.cels.anl.gov:argo/argopkgs.git
+```
+
+That local clone can be used to build components:
+```
+$ nix-build /path/to/argopkgs -A <packagename>
+```
+
+And install packages:
+```
+$ nix-env -f /path/to/argopkgs -iA <packagename>
+```
+
+It's reccomended you read `pkgs/default.nix` for a list of available packages.
+
+## usage (old way - use as a channel)
+
+You can also use `argopkgs` as a channel. It's not reccomended, because usage of `nix-env -i <packagename>` evaluates
+all derivations to get their `name` attribute - which in the case of argopkgs involves cloning a few repositories.
 
 ```
-nix-channel --add https://xgitlab.cels.anl.gov/argo/argopkgs/-/archive/master/argopkgs-master.tar.gz argopkgs
-nix-channel --update argopkgs
+$ nix-channel --add https://xgitlab.cels.anl.gov/argo/argopkgs/-/archive/master/argopkgs-master.tar.gz argopkgs
+$ nix-channel --update argopkgs
 ```
 
 which lets you install packages:
@@ -55,7 +74,7 @@ $ nix-shell ~/.nix-defexpr/channels/argopkgs -A candle
 if your key is already on the runners, you can redeploy them with:
 
 ```
-$ g clone git@xgitlab.cels.anl.gov:argo/argopkgs.git
+$ git clone git@xgitlab.cels.anl.gov:argo/argopkgs.git
 $ cd argopkgs/deployments
 $ nix-shell
 [nix-shell]$ ./deploy-runners.sh
@@ -63,7 +82,7 @@ $ nix-shell
 
 or reboot/restart their gitlab-runner daemon with:
 ```
-$ g clone git@xgitlab.cels.anl.gov:argo/argopkgs.git
+$ git clone git@xgitlab.cels.anl.gov:argo/argopkgs.git
 $ cd argopkgs/deployments
 $ ./restart-runners.sh
 $ ./reboot-runners.sh
@@ -75,8 +94,10 @@ these are deployed via NixOps:
 
 - defined in file [deployments/ci.nix](deployments/ci.nix)
 - deployment state is in  `deployments/argo.nixops`.
-- there are four shell scripts in [deployments/](deployments/) that can be used to redeploy, reboot, and restart systemd services.
-- access is granted to all keys in [deployment/keys](deployment/keys) so submit a pull request and ask someone to redeploy if you need access.
+- there are four shell scripts in [deployments/](deployments/) that can be used
+  to redeploy, reboot, and restart systemd services.
+- access is granted to all keys in [deployment/keys](deployment/keys) so submit
+  a pull request and ask someone to redeploy if you need access.
 
 The gitlab-runner service is called "gitlabrunner" on the machines. That
 systemd service registers runners on ExecStart and unregisters them on