Re: description of recovery mechanisms

[Glen Turner <glen.turner@aarnet.edu.au>]

Julian Satran wrote:
> None of the major TCP/IP protocols have it and they have all added
> details in a pragmatic fashion.

Black David replied:
> TCP error recovery is specified in excruciating detail down to
> minor tweaks in the retransmit algorithms.

That lack of detailed state diagrams for TCP notoriously
lead to an undefined state that needed to be clarified
in a later RFC.  There was then a lot of work, recorded
in the academic literature, to check that no other undefined
states existed.

And, as illustrated in earlier work in this list, we are
still suffering in a lack of precision in the definition
of TCP's Urgent field.

Glen