That's helpful feedback. Thanks, Joe. (For those who haven't, take a
quick look at section 3.8 of RFC 793.)

The milestone as-worded does not make reference to "abstract", but I
think there is understanding among the chairs and the authors that
this is implied; once adopted, the WG can come to consensus on what
constitutes an abstract API, and rework the document to meet this
definition. (We can also reword the milestone to specify "abstract"
explicitly, though I'm not sure that's really necessary.)

Kyle

On Wed, Apr 13, 2016 at 12:07 PM, Joe Touch <[email protected]> wrote:
>
>
> On 4/13/2016 7:20 AM, Kyle Rose wrote:
>>> Not in an RFC, IMO. Doing so sets details of the interface in stone,
>>> whether by declaration or implication.
>>>
>>> If you can't do it without that sort of example, IMO you have not
>>> succeeded in the "abstract" part of an abstract API.
>>
>> Is there an example you'd recommend of an RFC that does this the right way?
>
> The best one is one of the oldest - RFC793, but that focuses on data
> plane events and connection status, rather than parameters for
> extensions such as this.
>
> However, IMO, it's still a good reference.
>
> Joe

_______________________________________________
Tcpinc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tcpinc

Reply via email to