RE: Summing up on visibility(?)

> Actually, I didn't think of asking them to do anything.  I was just
> prying at some stuff that looked to me like popular fallacy, in the hope
> that if distinct points were put to rest, different conclusions might
> emerge.  I think I showed how a legacy service with no idempotent
> operations can be wrapped in an idempotent interface [1].  This was
> counter to a claim that the only way to get to idempotence was to
> burden the client with sequence numbers.  No one has answered my
> claim or provided a harder case.  This was supposed to be a reason
> why RM was necessary.  I'm just wondering if that opinion moved
> even slightly in response to my post.

Can you repeat the explanation?

From what I know if an operation is not idempotent than nothing can make it
idempotent. However, you can allow an invocation to be idempotent by
allowing multiple instances of the same invocation and detecting and
eliminating duplicates. The operation is still not idempotent (it has not
changed) but the interface allows idempotent invocation.

That's what RM achieves using sequence numbers. Another way is to use state
identifiers which are a more specific form of sequence numbers. But all
these are different implementations of the same model. What's the other
model you are proposing?

arkin

>
> If I were going to make a request of the WSA WG, I guess it would
> be to conduct a "fair trial" of the issue.  I'm not going to run after
> people
> and demand that they agree with me, unless you think I should. ;-)
>
>  > Mark
> > raises the "visibility" issue periodically as a principle that should
> > somehow be respected, and it appears that most of us don't get
> the point.
> > To the limited extent that I understand what you're getting at here, it
> > seems to me that XML supports "visibility" because 3rd party tools,
> > intermediaries, etc. can extract useful information for
> routing, cacheing,
> > security, etc. without truly "understanding" what's going on.
>
> When you say "supports" do you mean "enhances", or just "doesn't
> obstruct"?  Maybe that's the disconnect.  If the former, I suspect
> that "support" takes the form of elaborate configuration, in which case
> the burden is on the admin to "program" intricate sets of rules into
> the firewall config, based on text patterns.  Maybe XML, by keeping
> content ascii, eases that burden, but it's still a crusher, if you ask me.
>
> Walden
>
>

Received on Friday, 10 January 2003 16:49:54 UTC