On 28 Sep 2009, at 5:31 pm, John Cowan wrote: > That strikes me as over-specification. Most people in other languages > don't worry much about the range of the largest kind of fixnums they > have, > and it should not be a requirement that Scheme implementations signal > an error on fixnum overflow -- returning an inexact result should > still > be fine. > > But that does mean that it's acceptable to return either exact or > inexact > 2^24 when you add 2^23 to 2^23, which seems to count as meaning > "dependent > on an implementation-specified integer range".
Would it be acceptable to say: 1) R7RS compatible Schemes support exact integers up to some implemendation-defined maximum, which may depend on runtime environmental conditions (eg, limits of RAM) or some static limit (eg, limit of fixed-sized integers), but the time complexity of numerical operations may worsen as the numbers get bigger, and overflow results in an error being throw. 2) (as an SRFI) implementations may provide special sets of numerical operations with prefixes - fx+ fx- etc - that have special semantics: fx+ might work only in some defined exact integer ring and wrap around, f+ might return exact integers when the result is an exact integer within a range but return inexact reals when presented with rationals, inexact reals, or exact integers that are too large to fit in some bound (so would be good for real-time systems that still want as accurate maths as they can get), all with more tightly bound time complexities, for the real-timers? I agree with John's desire to provide specialised numerical semantics for special uses (real time systems being dear to my heart, in particular), but I worry vaguely about applications expecting different semantics of basic operations like + based on some big nasty global property of the implementation. So a small realtime embedded Scheme might provide +-*/ that check for overflow, and fx+ that work on simple integers, and f+ that does its best to be as accurate as it can while still using well-bounded time and space? ABS -- Alaric Snell-Pym Work: http://www.snell-systems.co.uk/ Play: http://www.snell-pym.org.uk/alaric/ Blog: http://www.snell-pym.org.uk/archives/author/alaric/ _______________________________________________ r6rs-discuss mailing list [email protected] http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
