W3C home > Mailing lists > Public > www-tag@w3.org > February 2009

Re: Last minute input to discussion re 'on the boundaries of content negotiation in the context of the Web of Data'

From: Dan Brickley <danbri@danbri.org>
Date: Fri, 20 Feb 2009 18:21:02 +0100
Message-ID: <499EE67E.702@danbri.org>
To: wangxiao@musc.edu
Cc: Michael Hausenblas <michael.hausenblas@deri.org>, "www-tag@w3.org" <www-tag@w3.org>, "timbl@w3.org" <timbl@w3.org>, Richard Cyganiak <richard@cyganiak.de>
On 20/2/09 17:52, Xiaoshu Wang wrote:
> <snip>
>> If it is an IR, we know we can preserve it by preserving the 0s and
>> 1s, alongside information aiding their interpretation.
> The "it" refers to the IR or its representation? I think it should be
> the latter because we can digital preserve something's state encoded in
> bits but not the thing itself. This is, what I think, one of the culprit
> of IR-definition, to encourage the practice of psychological
> identification of representation with resource as I have discussed in my
> manuscript[1].

Yes, I should have been clearer.

My idea was that an IR itself, is something that can be fully serialized.

On that story, a data file on my computer's USB hard drive can be an IR, 
while a thermometer attached to its USB port wouldn't be, although both 
could have HTTP representations.

I forgot to emphasise that an important distinction is that many/most of 
the serializations will be partial and lossy. A useful piece of metadata 
would be to know which (if any) of the representations are adequate for 
reconstructing the mysterious and hidden original.

> Philosophers are often criticized of hypostasizing or reifying things,
> that is, to create something for the purpose of one's theory. This is
> exactly what IR is doing. Nevertheless, we can follow WVO Quine's
> criteria. That is: hypostasis is O.K. as long as there is an ontological
> commitment. If TAG intends to define IR, define it in an ontology such
> as in RDF. At least in this way, people can derive some conclusion about
> it, other than an arbitrary term.

+1

If we must have a concept of IR, let's engineer a useful one, ie. choose 
definitions that correspond to the most practical and useful 
distinctions we can think of (hence the backups/archival suggestion 
above). I don't see the world as falling into two giant and obvious 
natural kinds - IRs and non IRs. Consequently, any such distinction is 
one we must very carefully invent, since there are various 
similar-but-different distinctions we might make. To date, much of the 
IR discussion has rested on appeals to obviousness...

cheers,

Dan

> [1]. http://dfdf.inesc-id.pt/misc/man/http.html
>
> Xiaoshu
>
Received on Friday, 20 February 2009 17:21:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:48:12 GMT