RE: Channel concept instead of simple Files

[Mike Griebling <grieblm@trt.allied.com>]

> >  Date: Wed, 10 Jan 96 08:08:27 EST
> >  From: Mike Griebling <grieblm@trt.allied.com>
> >  
> >  Let's do it the Oberon/F way.  I'm all for compatibility with an existing
> >  implementation.
> >  
> >  
> I like your enthusiasm.  But I checked the WinOberon/F version I got
> (1.0), and their Files module is basically the one that Sander
> presented.  O/F doesn't offer the kind of flexibility I'd like to see.
> Does anyone know if there is a more recent version of O/F with an
> enhanced IO concept?

There is a version 1.1 of OberonF.  Don't know if it has an enhanced IO
concept.
 
> The text Guy quoted previously has one good point: It identifies three
> levels of abstraction in the IO process.  
> 
>   chunk of data, identified by a locator
>       |  (eg file, memory block, pipe, tcp/ip)
>       |
>       |
>   basic reader/writer operating on bytes
>       | (offer SetPos, ReadByte, ReadBytes, WriteByte, WriteBytes)
>       |
>       |
>   reader/writer to translate high level data into byte stream
>         (eg. binary reader/writer, text rd/wr, see Sander's Files module)
> 
> This concept allows multiple readers and writers on one piece of data
> (just like the Oberon System does it on files), and allows to
> distinguish between the simple access layer that doesn't know a thing
> about the byte stream being read or written, and the interpretation
> layer that transforms some bytes into useful data.

This stuff quoted by Sander is the same as shown in my Files interface.
There are Directory, Reader, Writer, and some other type.
 
> -- Michael

Michael G.