Update README.md

README.md

@@ -7,24 +7,36 @@ Python programming courses using Jupyter or Colab notebooks.
 
 ## Who is this project for?
 
-The main target audience is teaching staff who develops programming courses
+The main target audience is teaching staff who develop programming or data science courses
 using Jupyter notebooks. The tools provided by this project facilitate addition
 of autogradable tests to programming assignments, automatic extraction of the
 autograding tests and student versions of the notebooks, and easy deployment of
 the autograding backend to the cloud.
 
-The main focus is Japanese universities, so the example assignments provided
-in the `exercises/` subdirectory are mostly in Japanese language.
+Initially, the project was started in collaboration with Japanese universities, so
+some of the example assignments provided in the `exercises/` subdirectory are in 
+the Japanese language. However, anyone using Jupyter notebooks (such as in Colab)
+can benefit from this project.
 
-## How to integrate autograder to your course
+## How to integrate the assistant to your course
 
-If you have a course based on Jupyter notebooks and want to integrate the
-autochecking tests, there are multiple different way how the autochecking tests
-can be run:
+There are two main ways to use the assistant:
 
-* Inside the student notebook (e.g. on Colab). The execution of autochecking
-  tests is handled within the same Python Runtime that student uses. Note that
-	this approach only supports self-checking, and cannot be used for _grading_
+* Student-driven self-checking with feedback. In this mode, the learner runs the tests interactively
+and receives immediate feedback about their code, instead of having to wait for an instructor or 
+teaching assistant to give them feedback. Although the tests themselves are not visible to the 
+learner, they can run them as many times as desried in order to learn the material, and the results
+from test runs are not recorded.
+
+* Grading. By running tests against submitted code and recording the results, the assistant can 
+also improve efficiency of instructor workload by automating the task of checking for specific
+code syntax and functionality.
+
+Depending your requirements for the above features, you can use either the "pure-Colab" 
+approach or the hosted approach.
+
+* Running the autochecking tests inside within the same Python Runtime that student uses. 
+Note that this approach only supports self-checking, and cannot be used for _grading_
 	student work. See the details in [docs/colab.md](docs/colab.md).
 
 * Hosted on Google Cloud Run. The scripts in this repository provide a server