At 17:26 97/08/20, John Whitley wrote:
>Perhaps what is needed are two tracks of language development,
>"Standard Haskell" and "Research Haskell". The research community
>continues to develop, distribute, and test new language concepts with
>less fear of disrupting existing users. After sufficient time the
>lessons from Research Haskell can be folded back into Standard
>Haskell.
This is a model that comes to my mind. It would answer the question posed
by John Whitley:
>-- What is the Ultimate Purpose of Haskell?
>What are the community's long-term goals, the collected hopes and
>dreams, for the language? In The Beginning, that seemed to be as a
>focus for research into functional programming languages. Haskell was
>born and has since matured greatly. Now the decision must be made as
>to its desired role in human computing endeavors.
There would be two objectives, one to continue the research development,
and another to standardize the features which are considered stable and
reliable.
The question is how to make the two Haskell variations communicate;
clearly the ideal would be to have one compiler that understands both.
Perhaps one could agree that code that is not specially marked is
considered to be "Standard Haskell", and "Research Haskell" code is name
space marked so.
Otherwise, the idea of marking code with version numbers is not new; it
is used in the LaTeX project (which sports a rather large number of
non-expert users). One would mark the code in the style "requires Haskell
1.4", with name, version, and optional version date (but one is not
required using this feature). It is like importing a module with a version
number attached to it.
Hans Aberg