New track for Dyalog APL

[rikedyp]

I posted in the general forum and was directed to the New Track documentation which has led me to post here.

I am interested in developing and maintaining an exercism track for Dyalog APL. Our organisation has begun work on GitHub - Dyalog/exercism: Repository to start an Exercism track for Dyalog APL. We are considering how tests should be written and will finalise this if we are approved to create an official track on exercism and can then begin work on the test runner.

[IsaacG]

Old post for reference:

Exercism track for Dyalog APL

I am interested in developing an Exercism track for Dyalog APL. Myself and another have started work on the repo structure here GitHub - Dyalog/exercism: Repository to start an Exercism track for Dyalog APL. I wonder whether there is process in place to transfer this to the exercism organisation or whether you would prefer to create a new repository and we simply commit the files into that? Then, should I create another thread to request a repository to implement the test runner?

Note, forum threads can be edited to change which category it is in.

[BethanyG]

Hi @rikedyp

Thanks for your posts, interest, and work thus far on a Dyalog APL track!

Jeremy (the person who usually reviews and does setup) is traveling, and won’t be available for the next 10-15 days.

In the meantime, it would be great if we could see your Hello World implementation and tests as mentioned by commenters in your other forum post.

[rikedyp]

We now have a hello-world exercise and a TESTS.md in the repository; unfortunately the forum will not allow me to post links to GitHub in the replies.

Dyalog/exercism/tree/main/exercises/practice/hello-world

Until the test runner works, users will need to import the exercise folder into Dyalog and run the tests as described in Dyalog/exercism/blob/main/docs/TESTS.md.

Can there be no submissions in the track until the test runner is working? Because presumably there is no other way to verify a correct solution.

[BethanyG]

Can there be no submissions in the track until the test runner is working? Because presumably there is no other way to verify a correct solution.

There would be no way to verify a correct solution on the platform. We’d be asking the students/coders to self-verify based on setting up the language and running the provided tests themselves locally. And that would then need to be supported.

So … no, we really don’t want to be doing that. It’s far better to take the time to get the test runner done and tested before track launch.

Besides, you will also need to get at least 20 exercises (with tests) made before launch as well - and those exercises can be used to work out kinks in the test runner.

And there is a bunch of other stuff on the site that doesn’t quite work if you take the test runner out of the process – like having verification of a passing solution in community solutions, and (I think) exercise locking/unlocking. There is more functionality I am probably forgetting.

[IsaacG]

@rikedyp

Dyalog/exercism/tree/main/exercises/practice/hello-world

exercism/exercises/practice/hello-world at main · Dyalog/exercism

[BNAndras]

@rikedyp

Until the test runner works, users will need to import the exercise folder into Dyalog and run the tests as described in Dyalog/exercism/blob/main/docs/TESTS.md.

Can there be no submissions in the track until the test runner is working? Because presumably there is no other way to verify a correct solution.

One could submit locally using the CLI and then mark the exercise as complete without a test runner. However, without an active test runner, someone could submit invalid Dyalog APL code and publish it. That significantly reduces the value of looking at community solutions if you don’t know whether the code presented actually solved the exercise.

Also, I haven’t seen stats on online editor vs. CLI submissions, but I would not be surprised if a good number of students use the online editor preferentially. Without a test runner, you’re not going to be as accessible to students, and that might drive away potential interest at a critical juncture. If you add a test runner after launch, those folks aren’t guaranteed to come back.

[IsaacG]

@rikedyp

Until the test runner works, users will need to import the exercise folder into Dyalog and run the tests as described in Dyalog/exercism/blob/main/docs/TESTS.md.

A large part of the platform is the ability to test the code directly from the website. Jeremy would need to weigh in on whether or not a track can go live without a test runner. A test runner might be a prerequisite, though Jeremy would need to weigh in when he’s back from traveling.

[ErikSchierboom]

We only launch tracks without a test runner when creating one is not possible. Is that the case here? If not, a test runner should be created. We have extensive documentation on building a test runner as well as building a track.

[ErikSchierboom]

Oh and I think an APL track would be a valuable addition, so consider it approved. Can you give me the github usernames of the people who will be building the track?

[rikedyp]

Thanks so much! A track runner will be possible and I’ll be looking at that in the coming weeks. @rikedyp and @jonocarroll are developing the track.

[ErikSchierboom]

I’ve just ran the bootstrap script so you should both have GH invites to join the Exercism org and then you’ll have access to the GitHub - exercism/dyalog-apl: Exercism exercises in Dyalog APL. and GitHub - exercism/dyalog-apl-test-runner repos.