Re: URI serialization issues

On 1/18/06, Kendall Clark <kendall@monkeyfist.com> wrote:
>
> On Jan 18, 2006, at 5:15 PM, Mark Baker wrote:
>
> > FWIW, I just offered that text as a starting point, and it probably
> > goes further than it needs to from my POV.  I think I'd be happy if
> > you simply "linked" the choice of GET and POST to the security section
> > in *some* manner.
>
> Well, yeah, that's a lot weaker than SHOULD'ing the link. But the WG
> folks who've responded (except for Pat) don't seem to see a
> connection here. I confess that I don't see it.

Ah, I just read some of the responses.  Dan's surprises me;

"Asking a question is asking a question; whether it's a hard question or
not seems orthogonal to safety."

If by "safety" Dan's thinking about the absence of a change in state
(which I think you mentioned was your understanding, Kendall), then I
disagree, but let me try to put my position in those terms.

If I asked you "What are the prime factors of this kazillion-digit
integer", that's a question that would take a large amount of time and
money to answer.  And if it costs money to answer, then that requires
your bank account be debited by that amount of money, which is a state
change.

Food for thought ...

Mark.