Skip to content

Latest commit

 

History

History
127 lines (92 loc) · 5.17 KB

INSTRUCTOR_NOTES.md

File metadata and controls

127 lines (92 loc) · 5.17 KB
Raw

Compilation of instructions and important things to remember for the different levels.

Level 1

We're going to be building an extremely simple web server using the 'wai' framework. Wai is a low level HTTP implementation similar to WSGI in Python or Rack in Ruby.

  • Hackage
    • Primary repository for Haskell packages and their documentation.
    • Package Index is extremely useful for discovering and interrogating packages.

Level 2

We're going to start building our own REST application by converting our "spec" into a data structure that will guide the construction of the rest of the application.

We'll also demonstrate the usefulness of newtype and how it can trivially make code more type safe, as well as providing more descriptive type signatures.

Level 3

Our application is starting to grow we should add another layer of assurance by writing some tests.

We will cover:

  • How to write a test-suite in the cabal file.
  • Usage of the hspec library as our test runner and how to feed it our application.
  • We will also write some tests using the hspec-wai package to give us some basic confidence in our error handling.

Level 4

This covers adding a database to our application.

  • We use SQLite because it requires the least amount of setup on the host machine.
  • All of the SQL is provided.

We will have the beginnings of some configuration in our application, but we will simply hardcode this for the time being as we will have a better solution in the next level.

We try to emphasise the distinction between data to be used externally in the Database vs the data we deal with internally in the application through the construction of our DBComment type.

The sqlite-simple-errors package is used to run our queries so that we have our errors as values, which is what we are doing in the rest of the application.

Again there will be a lot of documentation diving for this part as people will need to find the functions that will be used to run the queries against the database.

This level might be a bit hard to navigate as there are a bunch of files to walk through.

The steps for this level:

  1. src/Level04/DB/Types.hs
  2. src/Level04/Types.hs
  3. src/Level04/DB.hs
  4. src/Level04/Main.hs
  • Call out Traversable and Bifunctor typeclasses.
  • Call out the encoding instances & the automatic deriving of the ToJSON instances

Level 5

This is the "ExceptT" level.

After enduring some of the annoyance of manually handling the Either values in various ways. This level has the students implementing their own version of the ExceptT monad transformer.

Enough of the motivation for this abstraction has been introduced by this point and demonstrating how effectively ExceptT solves this particular problem will be a relief in this and later levels.

Level 6

We construct a proper method of handling our configuration.

This is a complex level due to having to handle the two configuration datatypes, plus the monoid instance. Additionally the configuration file is a JSON file that we manually extract information from using decoding functions from the aeson package. There will be documentation diving and possibly tricky functions involved.

I think it is important to not let people become hung up on the Monoid instance (it is partially completed already) additionally care needs to be taken that searching for the functions they need from the aeson library isn't too painful. Give hints or just tell them which functions.

The optparse-applicative library is used to handle the command line parsing but the students do not have to implement it themselves. It's it a gnarly library despite its usefullness, so it's there for people to look at and refer back to. But that's about it.

Exceptions will be discussed when working on the 'readObject' function to load the file into a JSON Value.

Level 7

This is "The ReaderT" level.

Students will be required to copy their completed versions of functions from previous levels that will then break in this level and need to be refactored. The most interesting work will be undertaken in the AppM module as they must reimplement their existing transformer with the new functionality.

There will also be a lot of time spent in the DB module as the DB functions no longer require the DB connection to be passed in manually.

Also there are functions in the FirstApp/Main module that will need to be updated to handle the new shenanigans.

General Notes - More to add.

  • readme in each level folder.
  • monoid instance - rehash single/multiple number of possible instances. Don't let people hang too long on this point.
  • how to find the documentation for the Header / ContentType
  • more instruction that lead people to hackage documentation for Text/ByteString etc
  • mention that import lists may need to be updated for the new types
  • some editors will need to jump in and out of the different levels (close, cd, re-open)
  • Mention that it's fine to use case statements for Either handling, we make it okay at the end.
  • Mention that creating modules is easy, useful, and very helpful.

IMPORTANT!

  • Stephen Diehl - What I Wish I Knew Learning Haskell
  • Ask students if they would prefer access to a prepared VM with the code & an editor or two.