SML vs Ocaml for ECMA script spec.

Daniel C. Wang danwang74 at gmail.com
Tue Oct 24 19:52:27 PDT 2006


I was suggesting that those who were interested in different 
specification techniques do it on their own time, and leave you guys 
alone. :) When the next rev comes out people can flog it out or perhaps 
have a process to include more detailed specs as clarifying 
informational addenda...  If I remember correctly RS5Rn is in that 
style. The prose is the "official" standard and the semantics was 
"informational". 

It might be good to have such parallel efforts, as it might help to 
debug the spec you guys are developing. It's my experience that you can 
never believe anything is done right, unless you do it two times in 
different ways.

In particular proving type-safety is best done with a small-steps 
semantics. However, I don't think you guys want to tackle that in the 
spec. :) But it would be good to have something closely in sync with the 
spec so those in the research community interested in doing such a thing 
would have a good base to work from.

Dave Herman wrote:
>
> Well, no-- the working group has decided on ML as the spec language 
> for very good reasons we've already gone into. We aren't going to 
> develop two semantics in parallel on a tight deadline. :)
>
> Dave
>




More information about the Es4-discuss mailing list