I'm sorry, my comments were directed toward record based api interfaces
which are typical of pdas, cell-phones, etc.; for which it's not clear, at
least to me, that it's appropriate to try to attach ports to their record
and field based api stubs in any scheme friendly way, as scheme has no
notion of indexed record based file/streams, although concede if it were
elegantly accomplishable it would be nice. (and do apologize for wasting
this list's time on topics unrelated to Unicode).

> From: "Bradd W. Szonye" <xxxxxx@szonye.com>
> Date: Fri, 13 Feb 2004 19:08:31 -0800
> To: srfi-52@srfi.schemers.org
> Subject: Re: Encodings.
> Resent-From: srfi-52@srfi.schemers.org
> Resent-Date: Sat, 14 Feb 2004 04:08:40 +0100 (NFT)
>
> Paul Schlie wrote:
>> Please take a look at their typical api's for data interface &
>> storage, you may change your mind.
>
> I am already quite familiar with the API for record-based text I/O on
> mainframe systems. You don't need to do anything special, just read and
> write. Cobol and C compilers do that just fine in that environment, but
> only because they abandon the "text is a filtered byte stream"
> abstraction that is totally inappropriate for those systems.
>
> This is an *old* issue, and the solutions are well-known. Your zeal for
> imposing an abstraction upon systems where it doesn't make sense is
> impressive but misguided.
> --
> Bradd W. Szonye
> http://www.szonye.com/bradd
>