[PL-sem-jr] Paper critique session

[Ming-Ho Yee]

Dear all,

We will be taking a break from Coq next week or the week after, to critique
and help improve Ben Chung's ECOOP paper for the camera-ready deadline.
Title and abstract are at the end of this email.

(Note that we still have our regular meeting tomorrow.)

If you'd like to provide feedback (and non-juniors are welcome!), please
fill out the following poll with your availability:
https://www.when2meet.com/?6918406-lXGKl. I'll follow up when we have a
room booked.

The plan is to read the paper and prepare feedback before the meeting. Ben
suggests one or more of the following:

1. a quick skim of the paper
2. focusing on sections 1, 2, 3 (short, tells a story)
3. focusing on sections 4, 5 (longer, more technical, also tells a story)
4. focusing on section 4 (main technical content)

At the meeting, we'll start with a quick discussion of the paper's content,
and then take turns to give feedback. Try to choose feedback that other
people can learn from, and save the low-level details for discussing with
Ben offline.

Thanks!
Ming-Ho

------


TITLE: KafKa: Gradual Typing for Objects
>
> ABSTRACT: The enduring popularity of dynamically typed languages has
> motivated
> research on gradual type systems to allow developers to annotate
> legacy dynamic code piecemeal. Type soundness for a program which contains
> a
> mixture of typed and untyped code cannot mean the traditional absence of
> errors. While some errors will be caught at type checking time, other
> errors
> will only be caught as the program executes. After a decade of research it
> there are still a number of competing approaches to providing gradual type
> support for object-oriented languages. We introduce a framework for
> comparing gradual type systems, combining a common source languages with
> KafKa, a core calculus for object-oriented gradual typing.  KafKa
> decouples the semantics of gradual typing from those of the source
> language. KafKa is strongly typed in order to highlight where dynamic
> operations are required.  We illustrate our approach by translating
> idealizations of four different gradually typed semantics into the core
> calculus and discuss the implications of their respective designs.
>
-------------- next part --------------
HTML attachment scrubbed and removed