Re: [tcpm] Re: [dccp] Re: DCCP and draft-ietf-v6ops-v6onbydefault-01.txt

[Fernando Gont <fernando@gont.com.ar>]

At 15:45 19/07/2004 +0200, Brian E Carpenter wrote:

> > Of course, if you provided application programmers with a higher-level 
> > API, all the soft errors and address cycling issues could be handled 
> > bellow the application.
>
> This is so kind of blindingly obvious that I'm surprised we are
> even discussing it as a matter of principle...
[....]
> Whether we can standardize such an abstraction, or only make
> recommendations, is another question.

That's precisely the point: don't make the Sockets API less abstract than 
it currently is, and don't bother programmers with stuff they shouldn't be 
required to know.
And propose the idea of a Higher-Level API, so that applications won't be 
influenced by changes in protocol specifics next time we're are faced with 
a protocol specific issue.

BTW, what do you think of the proposed solution?


--
Fernando Gont
e-mail: fernando@gont.com.ar || fgont@acm.org