W3C home > Mailing lists > Public > www-tag@w3.org > August 2007

RE: New Editors Draft of the httpRange-14 Finding

From: Rhys Lewis <rhys@volantis.com>
Date: Wed, 29 Aug 2007 00:53:07 -0700 (PDT)
To: "'Alan Ruttenberg'" <alanruttenberg@gmail.com>
Cc: "'www-tag'" <www-tag@w3.org>, "'Jonathan Rees'" <jar@creativecommons.org>
Message-ID: <005001c7ea11$fbffd6e0$0202fea9@volantisuk>

Hello Alan,  

> -----Original Message-----
> From: Alan Ruttenberg [mailto:alanruttenberg@gmail.com] 
> Sent: 24 August 2007 17:57
> To: Rhys Lewis
> Cc: 'www-tag'; 'Jonathan Rees'
> Subject: Re: New Editors Draft of the httpRange-14 Finding
> 
> 
> On Aug 24, 2007, at 3:53 AM, Rhys Lewis wrote:
> 
> > Hello Alan,
> >
> > My impression is that the language in AWWW is deliberately that way 
> > because there is no hard and fast rule about what defines 
> 'equivalent' 
> > representations. I suppose that we might say that authors make 
> > assertions about equivalence when they create multiple 
> representations 
> > and offer them via content negotiation.
> 
> It's ok if they make the assertions. But there needs to be a 
> way to see that they make sense (at least if the plan is that 
> the SW is to be used for science).
> 
But that's the core of the problem isn't it? We can encourage people to
make assertions that make sense, but there is no way to check it. At some
level, languages like RDF and OWL help authors make consistent assertions.
I'm forever checking my ontologies to make sure I haven't written complete
rubbish. But I can still assert that the moon is made of cheese if I wish.
RDF lets me make such assertions with precision, but not necessarily
accuracy.

I don't see that web technology can provide any more guarentee of
correctness of assertions than a printing process that allows papers to be
distributed in learned scientific journals. Any sense of correctness for
papers needs to be applied above that level, during the refereeing
process, for example.


> >  By the way, I'm not convinced that lossless transformation is the 
> > right model for this. For example, I believe that 
> representations that 
> > differ only in language are appropriate to serve via content 
> > negotiation. HTTP is explicitly set up to support language 
> as one of 
> > the criteria for such negotiation [1]. The HTTP specification notes 
> > that the reason it's called content negotiation and not format 
> > negotiation is precisely because representations are not 
> necessarily 
> > different formats of the same content.
> 
> I didn't offer it as the right definition, just as an example 
> of  *a* definition, which "representation" currently lacks.
> 
> >  Actually, there are some intersting use cases in 
> accessibility and in 
> > support for small mobile devices that are related. Suppose I mint a 
> > URI and assert that it identifies a resource that let's you find a 
> > barbeque I'm holding this weekend (chance would be a fine 
> thing with 
> > the summer we've had!)
> 
> That's a nice example. The resource is "a thing that lets you 
> find a barbeque". This measures the information content as 
> the outcome of the agent evaluating the instructions. Is that 
> a commonly accepted idea about what an information resource is?

I've not seen it written like that, though I've not been doing this very
long.

> 
> >  Anyhow, for users with traditional Web access, I provide a 
> > representation that includes a map, delivered as a large, colour 
> > image, with a big arrow showing the location. If I want to support 
> > users with visual disabilites, or people accessing my site with a 
> > small, text only mobile phone, this won't be much use. I might want 
> > them to receive directions to the event instead. I might 
> deliver them 
> > as text or even as an audio clip.
> 
> Are these the same information resource? Or different answers 
> to the same question?

That is exactly the question. I tried to answer it in the following para.
In
short, I think they can be if, as author, I make an appropriate assertion
about
the URI.

> 
> >  Now, the question is, can I legitimately serve the map, 
> the textual 
> > directions and the audio as representations of the same URI? If I 
> > claim that the URI identifies a map of the location of the 
> barbeque, 
> > then I think the answer is no. However, if I claim it identifies a 
> > resource that let's users find the barbeque, I suspect the answer 
> > might well be yes, though I would not be surprised to find that 
> > opinions differ.
> 
> I'm happy with any definition that is operational and can be 
> consistently applied.
> 
> >  In short, I think equivalence is in the eye of the author.
> 
> We're going to have a real communication problem on the 
> semantic web if we don't fix that.
> 
> >  On the specific question of whether a jpeg is a 
> representation of a 
> > person, I think the consensus is no. There has been a lot of 
> > discussion about related topics on this list over the last 
> couple of 
> > months. If you mean person the way I think you mean it, then that 
> > would be a non-information resource, as described in AWWW.
> > These have no representations and the current feeling is 
> that it would 
> > be misleading to serve a representation if the associated URI is 
> > accessed. The range-14 finding is about what should be returned if 
> > such an access occurs.
> 
> There is an issue that I'm betting will bite. You 303 to a CN 
> manage URI. There will be temptation to allow (to take your 
> example) image/ jpg return a picture in response to 
> http://www.example.com/ solar_system/Mars/moons
> 
> Is this allowed, along with the rdf and html versions? If so, 
> what is the IR that these are all representations of. If not, 
> how does one associate a picture to the non IR?

That's a really good question. I want to tease it apart a bit. Since the
303 says 
nothing about what you might expect to find, then with the current
definitions
it would be fine to serve an image from the returned URI. However, the
question
of whether you could serve HTML, RDF and a JPEG from that URI would depend

on what you asserted the URI to represent. That information might be
available
from some other source, such as text in an HTML link from some other page.
However, 
it's not available when following the 303.

The TAG is actually starting to think about the whole business of
redirection
and how to find out more about URIs without dereferencing them. There
should be 
some discussion at our next F2F in the middle of next month.
> 
> It would be good to explicitly clarify this in the note.
> 
> More comments forthcoming at some point.
> 
> -Alan
> 
> 
> >
> > Of course, you could content negotiate between a jpeg, gif, png... 
> > (pick your favourite image format) for a URI that you claim is a 
> > picture of a person.
> >
> > Best wishes
> > Rhys
> >
> > [1] http://www.ietf.org/rfc/rfc2616.txt (Section 12 Content
> > Negotiation)
> >
> > From: Alan Ruttenberg [mailto:alanruttenberg@gmail.com]
> > Sent: 24 August 2007 05:48
> > To: Rhys Lewis
> > Cc: www-tag; Jonathan Rees
> > Subject: Re: New Editors Draft of the httpRange-14 Finding
> >
> > Hello Rhys,
> >
> > The problem I have always had is knowing what the 
> conditions are for 
> > two representations to be of the same resource (and are the 
> resources 
> > information entities, or only the representations? - can I content 
> > negotiate for the jpeg "representation" of a person?)
> >
> > Using translations as an example is particularly 
> problematic, as most 
> > language translations are not exact because there are inevitably 
> > cultural attachments to the words that can not be easily 
> understood by 
> > non-native speakers.
> >
> > I can understand some much more constrained than what I 
> perceive to be 
> > extremely loose language the AWWW, and this document as 
> well, use. For 
> > instance we could say that two representations are of the same 
> > resource if there is a documented algorithm implementable by a 
> > computer that losslessly transforms one into the other, as 
> with a byte 
> > sequence and its gzip compressed version.
> >
> > My worry has been that unless there is some way for someone 
> to say:  
> > "No, you are wrong, these two things are *not* 
> representations of the 
> > the same thing" then the term "representation" is meaningless.
> >
> > Regards,
> > Alan
> >
> > On Aug 23, 2007, at 3:00 AM, Rhys Lewis wrote:
> >
> >> Hello everyone,
> >> Could I just take a moment to thank Roy and David for 
> their extensive 
> >> comments on the latest draft. They provide excellent input for the 
> >> forthcoming TAG face to face meeting next month.
> >> I hope to make progress on specific points before that meeting. If 
> >> so, I'll respond here on particular topics.
> >> Thanks again
> >> Best wishes
> >> Rhys
> >
> 
> 
Received on Wednesday, 29 August 2007 07:53:21 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:47 GMT