I received this message from John Reppy just now and thought it might be of interest to the Poly/ML list. In general I'm in favour of evolving libraries rather than the language so I'd tend to look favourably on this but I wondered if there were any views from users of Poly/ML.
David
-------- Original Message -------- Subject: [Sml-basis-discuss] Evolving the SML Basis Library Date: Mon, 24 Sep 2012 08:15:59 -0500 From: John Reppy jhr@cs.uchicago.edu To: Sml-basis-discuss@mailman.cs.uchicago.edu
The SML Basis Library specification was finalized and published about eight years ago. I think that it is past time to consider changes. I see these as falling into three groups:
1) incremental additions of operations to existing modules
2) addition of new signatures, structures, and functors
3) changes to the semantics (including APIs) of existing Basis Library modules
We need a process to manage these changes; something like the Scheme RFI (http://srfi.schemers.org) approach should work. We can host the proposals and specifications at Chicago. We also need agreement from maintainers of the major SML implementations that they are willing to support the process by implementing the changes.
I hope that this message will start a discussion about the process and provoke suggestions for improvements to the Basis Library (I've got my own list of incremental additions that I'll post soon).
- John
_______________________________________________ Sml-basis-discuss mailing list Sml-basis-discuss@mailman.cs.uchicago.edu https://mailman.cs.uchicago.edu/mailman/listinfo/sml-basis-discuss
Yes of course, provided the changes emphasise new capabilities. Tinkering with existing library functions (renaming in particular) must be kept to an absolute minimum. Larry Paulson
On 24 Sep 2012, at 14:36, David Matthews David.Matthews@prolingua.co.uk wrote:
I received this message from John Reppy just now and thought it might be of interest to the Poly/ML list. In general I'm in favour of evolving libraries rather than the language so I'd tend to look favourably on this but I wondered if there were any views from users of Poly/ML.