[Info-vax] An alternative history of computing

[Arne Vajhøj]

On 7/25/2021 9:05 PM, Simon Clubley wrote:
> On 2021-07-25, Arne Vajhøj <arne at vajhoej.dk> wrote:
>> On 7/25/2021 8:41 PM, Simon Clubley wrote:
>>>
>>> Those lower levels exist to provide services to the higher application
>>> levels. There is an expected set of core functionality at application
>>> level in order for a protocol to be useful and you can't call something
>>> open if you can't implement that expected set of functionality using the
>>> public standards.
>>
>> But you can implement whatever functionality you want.
>>
>> It may just not be compatible with somebody elses
>> proprietary implementation.
>>
>> Let us take TCP/IP and database access protocol. Let
>> us assume that Oracle's protocol is a closed while
>> PostgreSQL's protocol is open. Then you can implement
>> a PostgreSQL compatible database server or a PostgreSQL
>> client, but you can't do the same for Oracle.
>>
>> And that means absolutely nothing for whether
>> TCP/IP is open or not.
> 
> But they are not considered to be part of the core TCP/IP stack.
> 
> Everything that ships as part of a TCP/IP stack (including SSH,
> Telnet, FTP, SMTP, etc) has a public specification that goes
> with it.
> 
> However, not everything that ships as part of the DECnet stack
> on VMS has a public specification that goes with it.
> 
> (And don't forget in addition that the public CTERM specification
> is missing the VMS specific bits.)

Telnet, SSH, FTP, SMTP etc. are individual specifications - they
are not part of the specs for TCP and IP.

And they do not always come with TCP/IP.

They often do, because the specs are open and open source
implementations exist making them easy to deliver.

But being frequently bundled with does not make them
part of a protocol.

Arne