Chris Hanson wrote: >> Then, a strong argument for keeping the STRING type is that >> a mutable, homogenous, disjoint, arbitrary-size sequence of CHAR >> is a vital abstraction for common I/O operations. >> > > I don't see how "mutable" is vital here, unless you are referring to > implementation of I/O buffers. I don't see that buffering need be done > using the text abstraction; vectors of characters will work fine. > >
Homogeneity (and, consequently, disjointness) are desirable for efficient buffers and so that implementations have at least the option of representations that are directly useful with foreign functions and system calls. "Text" is an interesting but poorly elaborated abstraction in this discussion. Should it be permitted to include non-Unicode characters, as for example some Japanese would like? Should it include various kinds of character or region property? Markers? In any event, perhaps it is useful to begin conceiving of an eventual TEXT type as a non-disjoint abstract type whose API is given as some form of extensible generics. -t _______________________________________________ r6rs-discuss mailing list [email protected] http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
