[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

This page is part of the web mail archives of SRFI 77 from before July 7th, 2015. The new archives for SRFI 77 contain all messages, not just those from before July 7th, 2015.

*To*: "John.Cowan" <jcowan@xxxxxxxxxxxxxxxxx>*Subject*: Re: arithmetic issues*From*: Thomas Bushnell BSG <tb@xxxxxxxxxx>*Date*: Sun, 23 Oct 2005 12:39:17 -0700*Cc*: bear <bear@xxxxxxxxx>, Aubrey Jaffer <agj@xxxxxxxxxxxx>, srfi-77@xxxxxxxxxxxxxxxxx*Delivered-to*: srfi-77@xxxxxxxxxxxxxxxxx*In-reply-to*: <20051023160247.GC7728@NYCMJCOWA2> (John Cowan's message of "Sun, 23 Oct 2005 12:02:47 -0400")*References*: <20051021145326.816C11B77BB@xxxxxxxxxxxxxxxxxxxxx> <20051021155906.GC16464@NYCMJCOWA2> <Pine.LNX.4.58.0510210910130.18969@xxxxxxxxxxxxxx> <20051022020312.GB5632@NYCMJCOWA2> <87mzl2mhd3.fsf@xxxxxxxxxxxxxxxxx> <20051022023824.GD5632@NYCMJCOWA2> <87r7adv5ii.fsf@xxxxxxxxxxxxxxxxx> <20051023160247.GC7728@NYCMJCOWA2>*User-agent*: Gnus/5.110004 (No Gnus v0.4) Emacs/21.4 (gnu/linux)

"John.Cowan" <jcowan@xxxxxxxxxxxxxxxxx> writes: > Granted, Scheme vectors could be implemented as sparse arrays (and Scheme > strings as cords). For that matter, both could be implemented as lists, > given a magic first cell that makes them disjoint from Scheme lists. > And for that mattter, Scheme lists could be implemented as machine-level > vectors, provided you are willing to live with all the behind-the-scenes > copying that would be required. Numbers could be Church numerals, and so on. > > But if Scheme vectors don't have O(1) performance (actually O(log k) on > modern hardware) in a given implementation, users are likely to vote > with their feet. You seem to have a flat-footed way of thinking about these things. An implementation could use sparse arrays for long arrays with empty space and linear arrays for those which are compact. What is wrong with that? Are you saying people will vote with their feet, leaving an implementation which provides for a certain kind of array using a sparse implementation, and choosing instead an implementation which doesn't provide for that kind at all? Once again, you are thinking that a Scheme datatype must be wedded to one and only one implementation. Wrong! Thomas

**Follow-Ups**:**Re: arithmetic issues***From:*Marcin 'Qrczak' Kowalczyk

**References**:**arithmetic issues***From:*Aubrey Jaffer

**Re: arithmetic issues***From:*John.Cowan

**Re: arithmetic issues***From:*bear

**Re: arithmetic issues***From:*John.Cowan

**Re: arithmetic issues***From:*Thomas Bushnell BSG

**Re: arithmetic issues***From:*John.Cowan

**Re: arithmetic issues***From:*Thomas Bushnell BSG

**Re: arithmetic issues***From:*John.Cowan

- Prev by Date:
**Re: arithmetic issues** - Next by Date:
**Re: arithmetic issues** - Previous by thread:
**Re: arithmetic issues** - Next by thread:
**Re: arithmetic issues** - Index(es):