• To: <xml-dev@l...>
• Subject: Re: Getting rid of entities (was: misprocessing namespaces)
• From: "Rick Jelliffe" <ricko@a...>
• Date: Tue, 5 Feb 2002 23:55:34 +1100
• References: <001b01c1ae3d$23765440$78fe0250@pcukmka>

 From: "Michael Kay" <michael.h.kay@n...>
 
> > Not always. See Carl Malamud's excellent book* on why the overlayered
> > OSI Reference Model failed to get
> > successful/popular/excellent(?) implementations
> > while TCP/IP has taken over.
> 
> Actually, the protocol stack we use today (SOAP, XML, HTTP, TCP, IP, ...)
> can be seen as a highly successful implementation of the OSI Reference
> Model. It's the OSI protocols that failed, not the layered reference model.
> And the protocols failed not because there were too few implementations, but
> because there were too many (one for each computer manufacturer), whereas
> TCP/IP had a single portable implementation that was available free of
> charge.
 
I think Malamud's book challenges this view. From memory he thinks that
TCP/IP does not fit in with the OSI reference model, because IP cannot
be decoupled from TCP. 

Cheers
Rick Jelliffe

• References:
  • RE: Getting rid of entities (was: misprocessing namespaces)
    • From: "Michael Kay" <michael.h.kay@n...>

• Prev by Date: RE: Categories of Web Service messages: data-oriented vs action-oriented
• Next by Date: RE: Categories of Web Service messages: data-oriented vs action-oriented
• Previous by thread: RE: Getting rid of entities (was: misprocessing namespaces)
• Next by thread: A heavier-weight proposal for character entity definition
• Index(es):
  • Date
  • Thread