Re: network endpoints

On Apr 25, 2008, at 10:33 AM, noah_mendelsohn@us.ibm.com wrote:

> Stuart Williams wrote:
>
>>> [Jonathan Rees wrote:]
>>
>>> I'll make the change in a day or two or three if I hear no outcry.
>>
>> I guess consider this a little cry out :-)
>
> Sorry for not chiming in earlier, as I've been traveling.  Anyway, I'm
> afraid I agree with pretty much everything Stuart is saying here.   
> I think
> the system is best thought of in layers.  When you refer to  
> endpoint, I
> presume you mean something that has some sort of persistent existence
> across HTTP interactions.

Not really. I just mean "the web" in the role of dealing with some  
particular URI.

I just want a way to say what is obvious but not said yet, that  
"information resources" relate in some way to the web. Remember that  
IRs, as I've heard repeatedly confirmed, are abstract things that may  
be put on the web, but are not inherently part of it. So we have to  
perform an explicit step to say that the behavior of the web can be  
consistent with a URI "identifying" to the web the abstract document  
that it denotes. This is part of the HTTP spec and therefore  
something that has to be articulated about HTTP semantics. I accept  
Stuart's suggestion that we talk about an abstraction "the web" as  
the apparatus for handling requests without diving deeply into its  
nature. I was going to make a node for the class of things of the  
form "the web at a URI" in the diagram, and an arc in the diagram  
connecting that node to Abstract Document / IR signifying that if a  
URI U denotes an IR X, then the web needs to deliver values, at U,  
that represent X's state at the time the request was processed.

"Endpoint" seemed like a nice shorthand to describe the web  
specialized to a particular URI. But since "endpoint" is making both  
of you squirm, I will try to think of something else. Maybe "URI- 
specific web behavior".

I'll make a new version of the diagram, in any case.

If I'm still being unclear then it may be what I'm trying to do is  
state something so numbingly obvious that it is invisible and  
therefore, when stated, sounds controversial. But that what semantics  
is about.

Received on Friday, 25 April 2008 19:46:01 UTC