[r6rs-discuss] [Formal] Allow compilers to reject obvious violations

[Anton van Straaten]

[I am posting this as an individual member of the Scheme community.  I 
am not speaking for the R6RS editors.]

Arthur A. Gleckler wrote:
> We've already dropped one of the core features of the language, its  
> REPL, from the specification.  

I don't know that the REPL specification was so much dropped as 
acknowledged to be mythical.  That change does not seem to pose any risk 
of affecting "the prototyping-friendly nature of the language", since 
implementations with REPLs are unlikely to discontinue them.

> Let's not go any further and lose the  
> prototyping-friendly nature of the language.

The formal comment which started this thread proposed that 
"Implementations *may* reject a library or program prior to 
execution..."  As Kent later put it, "Then this reduces to a quality of 
implementation issue---users can pick the implementation that most often 
exhibits the behavior they like."  If adopted, that approach would put 
much of the discussion in the current subthread outside of the scope of 
R6RS.

However, it's not clear to me from the current subthread whether anyone 
is objecting to giving this much discretion to implementations, and 
instead would like R6RS to either require or forbid aggressive early 
error detection.  If there are such objections, I suggest that they be 
raised explicitly (apologies if they were and I missed them).

Anton