Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] View of /papers/modulespaper/dissertation/proposal-structure.txt
ViewVC logotype

View of /papers/modulespaper/dissertation/proposal-structure.txt

Parent Directory Parent Directory | Revision Log Revision Log


Revision 3564 - (download) (annotate)
Thu Sep 30 13:33:05 2010 UTC (8 years, 9 months ago) by dbm
File size: 1230 byte(s)
initial import
Proposal

Set candidacy exam for Fri 1/16 (Dave arrives on Wed)
Contrasts between Harper-Stone and Definition semantics (certain edge cases) One discrepancy is in the handling of nonuniform types (irregular recursion)

Def'n admits but doesn't require this: a t = ... a t ... a t ... ==> \a. s = ... s ... s ... (a simple type generator)
Harper-Stone restricts recursion to this regular recursion. 


Set regular writing time


Identify core problems
Methodology & approach
Background



In modules, components are named, but in MixML, they are not necessarily named. Module expressions such as link X = [int] with [3] are considered inconsistent by the reference implementation. 

Todo

(1) Make sure we identified all subtle points such as export-only projection limitation + binding workaround, asymmetric selection in recursive 

What does it mean to select out signature components? There are situations where projection from a incomplete module is well-defined: 
{t=[:0],x=[:int]}.x
{t=[:0],x=[:t]}.t

Also, the other way around is a little tricky:
m={l0=cm,l1=im} where cm is complete and im is not.
m.l0 is illegal
m.l1 is ok

Dreyer inspired by Bracha/Cook, Bracha/Lindstrom
What do C linkers do for cross-module linking?

root@smlnj-gforge.cs.uchicago.edu
ViewVC Help
Powered by ViewVC 1.0.0