On May 8, 2016 at 9:25:33 PM, Richard Eisenberg (e...@cis.upenn.edu) wrote: > > I do absolutely think we should be cautious about addressing unimplemented > behavior. > I would be strongly against a new type-system extension that hasn't been > field-tested. > However, I do think pondering lexical/parsical changes (such as limber > separators) > should be considered in scope.
While such changes should definitely be in scope, I do think that the proper mechanism would be to garner enough interest to get a patch into GHC (whether through discussion on the -prime list or elsewhere) and have an experimental implementation, even for syntax changes, before such proposals are considered ready for acceptance into a standard as such. But I also agree that discussion of things which may be in the pre-implementation phase are in scope — just that the next step is to get them past that phase before they’re considered for inclusion as such. There are enough traps in parsing that specification without an implementation is always a risky choice. Recall for example the case of fixity resolution, finally fixed in Haskell2010 (https://prime.haskell.org/wiki/FixityResolution) where the behavior in the report was at variance with all implementations. I would hope that if a segment of the prime committee wants to test-run an experimental syntax feature, then this would take sufficient precedence over concerns regarding “language fragmentation” that the GHC team would be open to guarding it behind a flag. Given the number of GHC developers involved in this effort, I think that’s not a bad estimation :-) Cheers, Gershom _______________________________________________ Haskell-prime mailing list Haskell-prime@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-prime
