Improve name resolution in Liquid Haskell #180
Conversation
facundominguez
force-pushed
the
fd/lh-name-resolution
branch
from
3855f4b to
3d6d676
Compare
facundominguez
changed the title
There was a problem hiding this comment.
I think the description does a good job at explaining the technical context. Is there any size limit? If there's a bit more room, it might make the proposal more attractive to include even just one sentence about the motivation for something like LiquidHaskell (something about ensuring high assurance/correctness for critical software in a non-obtrusive way, for example).
Also, I wonder why name resolution is currently done two times? Is it for ease of implementation, for historical reasons, or because of objective technical difficulties that the candidate will have to solve?
| A proposal should include tentative answers to the following questions plus any | ||
| others that the candidate considers relevant: | ||
|
|
||
| * What names should be accessible in the specifications of a given module? | ||
|
|
||
| * What should be the mechanism to bring names in scope? | ||
|
|
||
| * How should the system negotiate or resolve ambiguous names? | ||
|
|
||
| * Which design and/or implementation steps should be done to adjust Liquid | ||
| Haskell to the proposed mechanisms? | ||
|
|
||
| * How much effort would the previous steps take? |
There was a problem hiding this comment.
I find it a bit surprising to require a proposal. It looks like there is some pre-project acceptance criteria here (which I did not in see in existing proposals), whereas I think there isn't any. You are simply outlaying what the project will look like right? If that's the case, I would reword to make this look less intimidating.
There was a problem hiding this comment.
I reduced the intimidation factor. But since GSoC already provides the guidelines for proposals, maybe I should just keep the requirement of the final blog post.
There was a problem hiding this comment.
I should just keep the requirement of the final blog post.
Done.
facundominguez
force-pushed
the
fd/lh-name-resolution
branch
2 times, most recently
from
801ac2f to
610227b
Compare
Only the patience of the student :) I'll consider your suggestion.
I don't know for sure, but my guess is it is a historical reason. Meaning that different projects modified the code base without consolidated goals. I don't anticipate technical obstacles that justify the current state of affairs, though I admit perhaps we learn something more about it during the project. |
facundominguez
force-pushed
the
fd/lh-name-resolution
branch
from
610227b to
f0b6dce
Compare
facundominguez
force-pushed
the
fd/lh-name-resolution
branch
2 times, most recently
from
b96da2b to
7ea7a81
Compare
| The following aspects haven't been answered explicitly so far for the Liquid | ||
| Haskell scenario. |
There was a problem hiding this comment.
| The following aspects haven't been answered explicitly so far for the Liquid | |
| Haskell scenario. | |
| In other words, to attain this supplementary stretch goal, one will need to answer those questions: |
Rationale: I think that's what you mean
There was a problem hiding this comment.
Thanks @smelc! Adopted with edits.
facundominguez
force-pushed
the
fd/lh-name-resolution
branch
from
7ea7a81 to
1fe74e8
Compare
|
@aaronallen8455, I think this is good to merge unless you have suggestions. |
No description provided.