Re: Question on the boundaries of content negotiation in the context of the Web of Data

Xiaoshu, Jonathan, Stuart, All,
So, yeah, thanks for coming back to the original question ;)

I had the pleasure to learn already a lot by following the discussions so
far and additionally had a nice chat with DanC on #swig channel [1]
recently, basically boiling down to: I, as the authoritative party minting
the URI *alone* may claim two representations are sufficiently equal. If I
do this in a reasonable way and the 'community' out there buys into it, it
is fine. If I claim two representations equal that, for example, are
obviously not, I'll be not credible and hence people will not use my
stuff/not trust me with respect to this issue anymore.

So, the only question left now is: can we take TimBL's  opinion expressed in
various hallway/IRC chats, the essence of this discussion I tried to
summarise above a bit sloppy and some existing wording as of AWWW1 and
create a document (which in turn could feed into a finding or as input for
an AWWW2 ;)?



Dr. Michael Hausenblas
DERI - Digital Enterprise Research Institute
National University of Ireland, Lower Dangan,
Galway, Ireland, Europe
Tel. +353 91 495730

> From: Xiaoshu Wang <>
> Organization: Medical University of South Carolina
> Reply-To: <>
> Date: Mon, 16 Feb 2009 11:42:08 +0000
> To: Jonathan Rees <>
> Cc: Michael Hausenblas <>, Ian Davis
> <>, "" <>
> Subject: Re: Question on the boundaries of content negotiation in the context
> of the Web of Data
> Jonathan Rees wrote:
>> Michael,
>> I care about this too. We've been telling people that *all*
>> co-representations, even exotic ones like turtle and XRDS, need to
>> "say the same thing", without providing a piece of writing that
>> explains why and with what normative force. I'll get it onto the
>> agenda - or at least into the hopper for any discussion of TAG
>> priorities.
>> I think the procedure is that the topic gets onto a meeting agenda at
>> the chair's discretion, then at the meeting, after discussion, we
>> resolve to create an issue.
> I would be interested in an answer in this direction too.  I am guilty
> of diverting Michael's question off to a different one.  I think that I
> might have guessed too much on what the potential direction of his question.
> Thinking of Michael's specific question -- i.e., the definition of two
> representations are equivalent, I am, in fact, not optimistic that there
> can or should be one such definition or recommendation. To use OO
> programming language as an example. Most of them have defined at least
> two kind of *equivalence*.  One is based on object identity and the
> other a user-implemented method, such as java's equals().  This is
> reasonable because had a language defined a fixed semantics of the
> second equivalence, it will definitely hurts its robustness as a general
> purpose programming language.  For me, I have always thought that URI is
> the interface to the Web and awww:representation the implementation.
> Hence, I don't expect that the outcome of a potential discussion will be
> anything concrete, if any.
> Xiaoshu
>> Jonathan
>> On Fri, Feb 13, 2009 at 7:34 AM, Michael Hausenblas
>> <> wrote:
>>> TAG members, Ian,
>>> Honestly, I must admit that I'm a bit disappointed. I *thought* I've clearly
>>> articulated my question (and I still don't know what is missing from my side
>>> that I get clear answers - please let me know) and added a request for an
>>> issue here. In my understanding both were not addressed/answered in a
>>> satisfying way. I appreciate when a TAG member (thanks, Jonathan, this helps
>>> me feeling a bit less lost) states [1] :
>>> 'I'm as starved of citations on this subject as you are.'
>>> However, I'd really like to hear from a chair or whoever feels responsible
>>> if or how the TAG intents to address my question/issue.
>>> ...

Received on Tuesday, 17 February 2009 13:26:14 UTC