150PLD-Fall2019 / Reading

3 stars 0 forks source link

Eval Criteria Criticisms #296

Closed elif01 closed 4 years ago

elif01 commented 4 years ago

Software Ecosystem:

Too heavily focused on Haskell usage. If it’s good work that should suffice. Strong accept criteria: Not everything that’s created is meant to be integrated with other things. Perhaps better: “Is the software ecosystem appropriate for the end-user’s needs?” “I should be able to clone it, compile it, be able to understand it”

Software Engineering:

You dont have to have an API What is an API (people don’t seem to agree on a definition!) -- “All the ways the user can interact with your language” -- We are evaluating the code THEY WROTE — how are we gonna do that? is API the correct term here? Why do you have to justify library usage? More of interest: “Did they use the features of the host language to their advantage?”

Abstract Syntax and Internal Representations

Spelling out “important things” — we want to put in a point of reference —> Alternative : instead of saying “missing important things”, focusing on their justifications of what they have included and what they havent included.

Related Work

Using “at length” perhaps focuses on quantity over quality — under what context what is sufficient?

Clarity


Suggestion: terms are well defined for the target audience

Documentation and Usability

Codegen / Metaprogramming etc:

What if your language wont be needing these capabilities? Matt poses a controversial question in response: “If it doesn’t do anything at compile time is it a language or a runtime library?”

Evaluation

"Most" and "all" rather than 3

jjkv commented 4 years ago

edited per these notes