Re: ACTION-115: Note on proxy graph URI

On 12 Oct 2009, at 17:45, Chimezie Ogbuji wrote:

> Response to Kjetil / Andy re: proxies is below (I have made changes  
> to the
> editorial draft):
>
> On 10/11/09 3:59 PM, "Seaborne, Andy" <andy.seaborne@hp.com> wrote:
>> I don't like the term proxy because:
>>
>> A/ (minor) "proxies" in HTTP occur in a lot of places so it's a bit  
>> overloaded
>>
>> B/ Having two names for the same thing is just something that  
>> happens on the
>> web. Both names have the same status.
>>
>> It's the same graph, accessible through a different name.  Each  
>> name is as
>> valid as the other and one is not a second class name; "proxy", for  
>> me, sort
>> of implies it's not a first class name.
>
> Ok.  I was using the term proxy to tease out whether we were talking  
> about
> URI 'aliases' or service endpoints for the purpose of clarity in the
> protocol model.  It appears the use case we are talking about is  
> that of an
> alias.

I think that is maybe is an alias, but for what exactly?

There are cases where it's not something that can be retrieved by any  
other URI request — other than an explicit CONSTRUCT with no  
variables, which is cheating, you can build any RDF graph that way.

I think Andy may have made the point about it being an alternative  
syntax for a CONSTRUCT/INSERT etc. request, but I'm not sure and I  
don't want to put words in his mouth.

> In which case, I believe this settles the issue of whether this  
> induces good
> 'HTTP behavior', since presumably you can use all the verbs
> (PUT/POST/GET/DELETE) uniformly on these alias URIs as though you  
> were using
> the IRI of the graph directly. In addition, conditional GETs would  
> work as
> expected.

I would expect so.

> "Kjetil Kjernsmo" wrote
>> So, it means it is a URI Alias. While that carries some negative  
>> connotation,
> perhaps we should just use that term "endpoint URI alias", or  
> something...?
>
> Well (WRT the negative connotation of 'alias') - although Web arch  
> generally
> frowns on multiple URIs to identify the same resource it doesn't  
> forbid it,
> and it appears this use case is an exemplar of why URI aliases are  
> sometimes
> necessary.
>
> I have changed the editorial text and the image to use URI alias  
> rather than
> URI proxy.  Note I've also updated the editorial text with some of  
> your
> suggestions, but I didn't add text about at risk since I think the  
> existence
> of editorial text that highlights the concerns and the fact that it  
> is a
> FPWD adequately addresses this.

I'm not a huge fan of the term "alias" here — it's technically true, I  
suspect, but probably misleading.

- Steve

Received on Tuesday, 13 October 2009 10:01:45 UTC