W3C home > Mailing lists > Public > public-awwsw@w3.org > May 2008

Re: network endpoints

From: Pat Hayes <phayes@ihmc.us>
Date: Thu, 1 May 2008 02:45:18 +0200
Message-Id: <p0623090cc43ec19a0a41@[]>
To: noah_mendelsohn@us.ibm.com
Cc: Jonathan Rees <jar@creativecommons.org>, "public-awwsw@w3.org" <public-awwsw@w3.org>, "Williams, Stuart (HP Labs, Bristol)" <skw@hp.com>, Tim Berners-Lee <timbl@w3.org>

At 6:34 PM -0400 4/30/08, noah_mendelsohn@us.ibm.com wrote:
>Jonathan Rees wrote:
>>  I might have talked about servers, but was discouraged from
>>  doing so by Stuart and Noah.
>It was not my intention to discourage you from talking about servers,
>though I'm not yet convinced that they're an important abstraction in
>telling the basic HTTP story.  Maybe they are.  I was discouraging the
>introduction of the term "endpoint", which I understood (perhaps
>incorrectly) to be an abstraction for something that's not quite the
>server or the resource, but is also persistent in the sense that, unlike
>representation, it exists outside the context of one particular HTTP

My understanding of 'endpoint' says that it can't be a server because 
one server can... what... host? contain? look after? handle? ... a 
number of distinct endpoints. The IHMC server handles a large number 
of different Web pages, for example.

>  > An IR defined to be one whose awww:representations convey the
>>  declaration of US independence isn't real in the same way since
>>  it would have those
>>  awww:representations regardless of what happened on the web.
>I'm not quite sure that the representations follow directly from the fact
>that the resource is the Declaration of Independence, but if there's an
>abstraction missing in the story I'm inclined to think that it may be the
>"assignment authority" or if you like the "owner" of the resource.  Let's
>say that I, Noah, have registered noahmendelsohn.com (I have) and that I
>choose to assign the URI
>http://noahmendelsohn.com/declarationofindependence.  I decide that the IR
>that I choose for this to represent is the text of the Declaration of
>Independence, including the lexical distinction of the front matter, the
>paragraphs, the signatures and so on.  I think I still have some freedom
>to decide which of the many (actually infinite) number representations of
>this information I will chose to serve.  For example, I could use large
>fonts or small, or a different type face.  I could send a text/html
>document and/or text/plain, etc.  All that's required is that the receiver
>be able to answer to come to agreement with me on the aspects of the
>representation I care about.  Crucially, it's not in general possible for
>a user retrieving a representation to determine whether the font size, for
>example, is something that I consider fundamental to the information
>content of the resource, or just an artifact of the representation I have
>chosen.  We can either model that by saying )a)that the resource is just
>the declaration, and that the particular representations served are
>determined by the responsible authority or (b) we can say that the
>resource is not in fact the Declaration in the abstract, but rather the
>particular representations, including choice of media types font
>specifications, etc.
>Either way, I don't find "server" or "endpoint" to be the missing
>abstraction; "resource authority" or if you prefer "assignment authority"
>might be something we have to talk about.

I find all this usage of 'authority' language very mysterious and 
hard to connect to anything real. Are we talking about architecture 
here? Because architecture knows nothing about things like authority 
and delegation. I'd suggest that we banish such talk, as it only 
confuses things. The very idea of 'ownership' doesn't seem to have a 
real place in this entire story, IMO. (I know this goes against 
deeply rooted ideas of the Web, but I have never found them to be 
anything but a kind of comforting fantasy.)


>If the authority chooses to use
>the mechanisms of some server software to create or choose the
>representations, so be it, but that's valid only because he/she chooses to
>delegate that responsibility.
>Noah Mendelsohn
>IBM Corporation
>One Rogers Street
>Cambridge, MA 02142

IHMC		(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32502			(850)291 0667    cell
http://www.ihmc.us/users/phayes      phayesAT-SIGNihmc.us
Received on Thursday, 1 May 2008 00:45:57 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:21:06 UTC