[r6rs-discuss] [Formal] Library form should prefix libraries

[John Cowan]

Michael Sperber scripsit:

> There's been a lot of discussion of this on the ML for SRFI 83,
> starting here:
> 
> http://srfi.schemers.org/srfi-83/mail-archive/msg00002.html

Thanks.  I've now read that thread in extenso, but it still seems
to me that if Scheme can cope with undefined delimitation of top-level
programs, it can cope with undefined delimitation of libraries.

> I'm not sure what you mean by "more convenient for both generation and
> processing".  The self-delimiting nature of `library' and the clear
> separation of the import/export section from the code body certainly
> makes it very easy to process for me.

I meant that when creating or reading a library in source form from
within Scheme, the text of the library must be written as a single
datum and read as a single datum.  Unless one writes the outer ( and )
using character-level I/O, it is not possible to separately generate the
parts of a library and write them sequentially to a stream; the entire
library must be built up as a Scheme object, which may be undesirable.
Likewise, unless one consumes the outer ( and ) and any surrounding
cruft (comments, etc.) with character I/O, it is impossible to process a
library streamwise: one must ingest the entire library as a Scheme object.

-- 
Values of beeta will give rise to dom!          John Cowan
(5th/6th edition 'mv' said this if you tried    http://www.ccil.org/~cowan
to rename '.' or '..' entries; see              cowan at ccil.org
http://cm.bell-labs.com/cm/cs/who/dmr/odd.html)