Re: [tap] [offlist] Re: Parse error vs failure

* Michael G Schwern <schwern@pobox.com> [2009-02-01 00:05]:
> Aristotle Pagaltzis wrote:
> > Basically. I think we should differentiate between
> > syntactical and semantic errors.
>
> Assuming a semantic error is treated as a failure, I'll take
> that as a yes.

Yes. :-)

I just didn’t quite like the “failure-with-warning” term: it
feels a bit too prescriptive. In practice most processors will
treat semantic errors that way, but I’d prefer to do the same as
the Atom WG, where we said we just tell you *exactly* what the
bits mean but what to do with them is up to you.

Although I guess “failure” can be seen as merely a meaning for
the stream as a whole, in which case “failure with warning” would
be another, in which case this all would be mere nitpicking. I
don’t care about the precise terminology though, as long as the
principle is kept in mind: don’t prescribe processor behaviour,
just describe document meaning.

Regards,
-- 
Aristotle Pagaltzis // <http://plasmasturm.org/>
_______________________________________________
tap mailing list
tap@ietf.org
https://www.ietf.org/mailman/listinfo/tap