Re: [webfinger] Question about device info discovery example.

[nov matake <[e-mail]>]

Cleaner :)

On 2012/12/22, at 14:48, "Paul E. Jones" <[e-mail]> wrote:

> I’m not. :-)
>  
> Wordsmithing… how about this:
>  
> Note that the host queried in this example is different than for the acct URI example, since the URI refers to a different host.  Either this host would provide a response, or it would redirect the client to another host (e.g., redirect back to example.com).  Either way, ...
>  
> I’m trying to make it clearer. I think the last version was not so clear.
>  
> Paul
>  
> From: nov matake [mailto:[e-mail]] 
> Sent: Saturday, December 22, 2012 12:44 AM
> To: Paul E. Jones
> Cc: [e-mail]
> Subject: Re: [webfinger] Question about device info discovery example.
>  
> Thanks, I'm happy with the change :-)
>  
> On 2012/12/22, at 14:39, "Paul E. Jones" <[e-mail]> wrote:
> 
> 
> Nov,
>  
> As I quickly looked at the text, I think we really should be consistent in using the hostname that immediately follows http(s):// or device:.  So, I’m going to change the WF draft accordingly.
>  
> So, the end of section 3.1 now looks like this:
>  
> An alias is a URI that is different from the “subject” URI that identifies the same entity.  In the above example, there is one “http” alias returned, though there might have been more than one.  Had the “http:” URI shown as an alias been used to query for information about Bob, the query would have appeared as:
>  
>   GET /.well-known/webfinger?
>            resource=http%3A%2F%2Fwww.example.com%2F~bob%2F HTTP/1.1
>   Host: www.example.com
>  
> Note that the host queried changed in this example, since the URI refers to a different host.  Either this host would provide a response, or it would redirect the client to another host (e.g., example.com).  Either way, the response would have been substantially the same, with the subject and alias information changed as necessary.  Other information, such as the expiration time might also change, but the set of link relations and properties would be the same with either response.
>  
> I also put “p1.example.com” as the host for the device: URI.  I do think the particular host to query is something we’ll have to take up on a URI-by-URI basis, but a very logical assumption with no other guidance is to query the host that is in the “hostname” position of the URI (if there is one).
>  
> For URIs like “tel”, the WF client would just have to be provisioned to know where to go query.
>  
> Paul
>  
> From: [e-mail] [mailto:[e-mail]] On Behalf Of Paul E. Jones
> Sent: Saturday, December 22, 2012 12:26 AM
> To: 'nov matake'
> Cc: [e-mail]
> Subject: Re: [webfinger] Question about device info discovery example.
>  
> Nov,
>  
> The host to query for the acct URI should be the host portion following the @.  For example, if one’s email ID is “[e-mail]” then the host to be queried would be nc.example.com.  That would be true for mailto, also.
>  
> For http(s): URIs like “http://www.example.com/some/path/”, there are one of two possibilities.  In the examples, I’ve suggested that to one would query the host “example.com”.  However, it probably isn’t the right place.  In the case of an http: URI, perhaps queries should be directed to the host itself (in this case www.example.com).  We need to answer that question: do we direct queries to the parent “host” (as I have it in the examples) or to the named host in the URI?
>  
> Paul
>  
> From: nov matake [mailto:[e-mail]] 
> Sent: Saturday, December 22, 2012 12:10 AM
> To: Paul E. Jones
> Cc: [e-mail]
> Subject: Re: [webfinger] Question about device info discovery example.
>  
> OK,
>  
> Then basically webfinger clients MUST use the host of resource parameter (at least when using acct, mailto, http and https schema),
> but extension spec MAY define another rule and in that case, client MUST follow the extension's rule.
>  
> Is it right?
>  
> Or webfinger itself doesn't specify anything for host usage?
> (It just define path "/.well-known/webfinger" and schema "https", but not host?)
>  
> On 2012/12/22, at 13:57, Paul E. Jones <[e-mail]> wrote:
>  
> 
> Nov,
>  
> Keep in mind that this is an entirely fictional example.  There isn’t even a URI scheme called “device”.
>  
> The answer to your question would come from a document (if it existed, but does not) that describes the “device” URI and how to use it within the context of WebFinger.
>  
> My thinking when I wrote this is that there are named devices on the network and one would query the parent domain to learn about the device.  The parent in this case beingexample.com.  It might be that there is a designated host that serves as the device info server called “devices.example.com” to which all queries are directed.  In any case, this is completely unspecified and the example is there only to illustrate the range of use for WebFinger.  It should not be viewed in any way as proper usage, though.
>  
> Paul
>  
> From: [e-mail] [mailto:[e-mail]] On Behalf Of nov matake
> Sent: Friday, December 21, 2012 11:40 PM
> To: [e-mail]
> Subject: [webfinger] Question about device info discovery example.
>  
> Hi all,
> 
> I'm new to this ML.
> 
> I'm writing a ruby webfinger library.
> https://rubygems.org/gems/webfinger
> 
> After reading usecases in section 3, I have a question.
> 
> In the device info discovery example in section 3.4, resource is "device:p1.example.com" but the host of discovery endpoint is "example.com", not "p1.example.com".
> Is there any reason why these hosts are different?
> How did the client decide the host of the discovery endpoint from resource URI?
> 
> Thanks
> 
> Nov Matake