[Info-vax] Rethinking DECNET ?

[Johnny Billquist]

On 2014-09-03 06:23, glen herrmannsfeldt wrote:
> Johnny Billquist <bqt at softjar.se> wrote:
>> On 2014-09-02 17:44, JF Mezei wrote:
>>> On 14-09-02 11:14, Bill Gunshannon wrote:
>
>>>> TCP/IP has never "pretended" to be anything OSI.  Some people have
>>>> tried to shoehorn it into the 7-layer Model
>
>>> The 7 layer model is not really specific to "OSI". It is a generic
>>> breakdown of networking functionality which was used in building the OSI
>>> networking stack, allowing each vendor to test compatibility of each layer.
>
>> No. The 7 layer model is specifically the OSI model. There is no
>> universal rule that you have to have 7 layers, or what should be in each
>> one of them. When we talk about the "7 layer model" we are explicitly
>> talking about the OSI model.
>
> As I remember the story, there was disagreement with some wanting
> eight and some six, so they compromised on seven.
>
> The idea of layers is more general, and the lower layers tend
> to be pretty obvious. The higher ones not so obvious.

Right. So if we just want to talk about a network stack as in layers, 
that is pretty universal and true. But exactly how many layers, and what 
to put in each one, is where people disagree.

Yes, the 7 compromise was, as far as I can remember based on americans 
generally favoring 6 layers while some europeans wanted 8. And in the 
end, they were not entirely clear what to actually even have in all of 
those layers... And the "OSI 7 layer model" is specifically the OSI 7 
layer model, since it deals with exactly how many layers you were 
supposed to have, and what to place in each layer.

	Johnny

-- 
Johnny Billquist                  || "I'm on a bus
                                   ||  on a psychedelic trip
email: bqt at softjar.se             ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol