W3C home > Mailing lists > Public > www-tag@w3.org > November 2008

Re: [XRI] Cooler XRI

From: Xiaoshu Wang <wangxiao@musc.edu>
Date: Wed, 19 Nov 2008 11:55:53 +0000
Message-ID: <4923FEC9.4070706@musc.edu>
To: Pat Hayes <phayes@ihmc.us>
CC: John Bradley <jbradley@wingaa.com>, "www-tag@w3.org" <www-tag@w3.org>

Due to some mess-up of mail servers, I have been late to join the 
discussion.  I have submitted a manuscript, titled "Exploring HTTP 
Content Negotiation", to WWW09, where the discussion is very relevant to 
this thread.  The very problem comes down on how we think about the Web. 

I don't know if it is appropriate to attached my manuscript in this mail 
so I won't. But if you want to read it, send me an email.  Here is some 
of my comments of the thread.

Pat Hayes wrote:
>
> On Oct 24, 2008, at 6:54 AM, John Bradley wrote:
>
>>
>> We all now understand that:
>>
>> * If an "http" resource responds to a GET request with a 2xx 
>> response, then the resource identified by that URI is an information 
>> resource;
>> * If an "http" resource responds to a GET request with a 303 (See 
>> Other) response, then the resource identified by that URI could be 
>> any resource;
>> * If an "http" resource responds to a GET request with a 4xx (error) 
>> response, then the nature of the resource is unknown.
>>
>> Givin this understanding the "Cool URI" document shows us how to 
>> construct URI for "real-world objects or things".
>>
>> As it happens every XRI is also about a "thing".  We have attempted 
>> to come up with better language than "thing".
>>
>> I quite like the description used by Stuart Williams of "Platonic 
>> ideal" in that when I the XRI =jbradley to refer to me I am not 
>> literally refering to me but to an ideal of me that can be descibed 
>> by meta-data in the same way that a mathmatician describes a circle 
>> via a formula.
>> Both myself and any phisical circle are crude appoximations of the ideal.
>
> I would beg you to reconsider this idea. You have just stepped into a 
> philosophical trench so deep that only a few specialized submersibles 
> can plumb its depths, and survival without specialized aids is 
> impossible. Unless you are prepared to deal with the elaborate and 
> expensive equipment needed (a graduate degree and an established 
> publication record in philosophical metaphysics is an absolute 
> /minimum/) I would strongly urge you to stick to the innocuous 
> terminology of "thing". This is not language that /needs/ to be 
> improved: its very banality is its power, as what is wanted at this 
> point is a term for something that can be, literally, anything. (The 
> philosophical word "entity" would have more resonance, but it has 
> already been co-opted by XML for a completely different purpose.)
It is true that this may put our foot into some hard philosophical 
issues.  But philosophy is not something only for the philosophers.  We 
all have our  philosophy, it is what guides us our actions, consciously 
or subconsciously. 
> All that actually matters for Web architectural discussions, in any 
> case, is that there are some, er, things that cannot be sent over a 
> byte stream, and that some names (URIs, XRIs, whatever) are required 
> to denote such things.
But to distinguish the "can" from the "cannot" needs philosophical 
guidance.  It requires an understanding of the relationship between 
"symbol", "reference" and "object".  It also requires to distinguish 
"information" from "meaning". 
> Further philosophical stipulation about the nature of these things is 
> a topic that belongs to philosophy rather than to Web science. And to 
> repeat my appeal, please, please do not engage in philosophy unless 
> you are seriously prepared for the resulting debates. It is not a game 
> for amateurs to play.
But philosophy can help.  In [1] (p. ix), Guy Fitzgerald said: one of 
the most important contribution that philosophy can do to information 
science is to highlight the various assumptions that underlies our 
action.  Our current thinking does have an implicit (but wrong) 
assumption that, I think, underlies the definition of "information 
resource" and "httpRange-14". 

I am an amateur.  Philosophy to me is pragmatic.  Whether something is a 
"Platonic Ideal" or "Resource' or "Thing" is irrelevant with regard to 
how the words are originally used.  What matters to me is whether a 
conceptualization of them can help me to clear my mind and do useful 
things. Reading some philosophical books in recent past have at least 
make my mind clear about what the Web and what the Semantic Web is. 
Hence, if I were to debate about my standing, it won't be about the 
meaning of the word but its consequence.   

>
> Allow me to point out some of the issues that immediately arise, but 
> would not have done if you had not tried to invoke the ghost of Plato. 
>
>> Givin that the XRI =jbradley names the PI(Platonic ideal) me how do I 
>> use that as a URI?
>>
>> I can use the proposed sub scheme for a http: XRI and put the 
>> relative XRI on a base http: URI:
>>
>> http://xri.net/=jbradley
>>
>> Now if this URI returns a 303 and link header information about where 
>> to retreve meta data about =jbradley and perhaps alternate resources 
>> relating to =jbradley based on content negotiation it is by the W3C's 
>> definition  a cool URI.
>
> True, but it refers to a Platonic Ideal. I, personally, would have no 
> idea how to even begin composing metadata about a Platonic Ideal of a 
> particular person. What is there to say about such a thing? One could 
> not possibly use this in FOAF, for example, or with any properties 
> from Dublin Core, since Ideals don't have friends or engage in authorship.
A particular person, as everything-else, *is* a "Platonic Ideal" (or 
Thing).   Not that there is a Platonic Ideal *of* a particular person.  
Of course, we could compose data (because I don't know what is a  
metadata?) of a particular person, which can have friends and engage 
authorship.

As I have pointed out in my manuscript, resource (Thing) is all about 
meaning. But meaning is not the same as information - a point clearly 
pointed out in Fred Dretske's book[2].  Once we understand this 
distinction, a lot of debates in the Web is clear (at least to me). 

For me, the essence of XRI is not about its *extensibility* or 
*resolution* protocol.  It is about its intent to describe the symbol 
(URI) but not its referent (Resource).  To put it plainly, what XRI 
intends is a framework where URI is composed of URIs.  However, the 
current URI is syntactically incomplete because its referential realm 
covers anything but URI.  This is what makes the debate of XRI so hard 
and long.  There are two solutions that are not mutually exclusive.  One 
is to use content negotiation and the other is a syntactic notation.  
The latter approach might require a modification of current URI spec if 
to be standardized.  I have discussed them in my manuscript.

The irony, to John and other XRI TC members, is that, if you take my 
analysis, many of the assumed feature of XRI, is unwarranted.  But the 
delight is that you can take the spirit of XRI and develop a framework, 
using content-negotiation, that works within the current web 
protocols.   Hence, you don't need TAG's approval to pass your resolution.

That is my two cents about the subject.

[1]. Fitzgerald, G. Forward. in Russel Winder, S.K.P., Ian A. Beeson ed. 
Philosophical Aspects of Information Systems, Taylar & Francis, (1997).
[2]. Dretske, F.I. Knowledge and the Flow of Information. MIT Press, 
Cambridge, MA (1981)

Xiaoshu
Received on Wednesday, 19 November 2008 11:56:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:48:08 GMT