On Feb 15, 2010, at 5:06 PM, Markus Karg wrote:
>>> It would have been more fair to tell Mr Fielding the whole story:
>>>
>> I thought I asked the question in a fair and non-predudicial way. I
>> think the terms "RPC-style", "fancy HTTP header" etc are quite the
>> opposite.
>
> The problem is that the question was so abstract that the answer would not
> deal with the problems of fancy headers and RPC, so it was not necessarily
> clear to him what we *actually* like to know: Whether *any* headers are
> allowed (which you cleared later on) or whether RPC-style URIs are allowed
> (which you didn't mention).
>
> If we want to get a clear answer to our actual question, we have to provide
> actual information what we are talking about. Never ever doubted what you
> actually asked in your first question, and you seem to notice it when you
> asked your second.
>
I think the question is fairly simple. You claim that HATEOAS when applied to HTTP does not include HTTP headers, others think differently. RPC-ness (which FWIW, I don't think any of us is in favor of) is orthogonal to that basic question.
Representation *is* an abstract term, we want to now how that abstract term maps to concrete HTTP components like headers and entity bodies.
Marc.