Re: Should the HTTP protocol become "inapropriate"

Hi David.

I think we are saying the same thing.

Perhaps a semantic difference in our use of significant vs specific.

On 30-Jul-08, at 10:04 AM, Booth, David (HP Software - Boston) wrote:

>> From: John Bradley [mailto:john.bradley@wingaa.com]
> [ . .  . ]
>>>>          * use the least specific bits of the domain name
>>>
>>>    That would also be fine if the owner of that URI space
>>>   chooses to establish that convention.
>>
>> This is convenient in that it allows multiple proxies without
>> registration.  The downside is that it confers special
>> meaning on a host name and can have an impact on people who
>> don't want to participate in the sub-scheme.
>>
>> Why should someone tell me that xri or xmpp is a reserved
>> host name.   I can see this leading to all sorts of great.
>>
>> It is equivalent to saying that part of the http protocol
>> will only apply to hosts named www.
>
> Hold on, it sounds like I misunderstood your short-hand.  When you  
> said one of the options was:
>
Sorry was referring to least significant , most specific or prefix as  
in the host name.

An example http://www.*.*/

My point is I think yours that having a host name of www should not in  
itself confer special meaning to the URI.

While tempting from a distributed configuration point of view,  it is  
bad design.

>>>>          * use the least specific bits of the domain name
>
> I thought you were talking about using http://hxri.xri.net/ as a URI  
> prefix to indicate that all URIs matching http://hxri.xri.net/*  
> conform to a particular URI convention and should be interpreted as  
> HXRIs.   It would *not* be okay to claim that all URIs matching http://hxri 
> .*.*/* should be interpreted as HXRIs.  The XRI TC does not have the  
> authority to define the special conventions for http://hxri.*.*/*,  
> but the owner of xri.net *does* have the authority to define special  
> conventions for http://hxri.xri.net/* .
>
>
I am arguing that the most significant, least specific or suffix part  
of the domain name be used.

an example http://*.*.xri./

This will require some work at the DNS level to allow auto  
configuration of multiple proxies.

I agree with you it is the better design.

I don't think using a single fully qualified domain name is desirable  
for scalability reasons.
I don't think it should be precluded,  but the design should support  
multiple proxies under different administrative control.

Regards
John Bradley
OASIS IDTRUST-SC
http://xri.net/=jbradley
五里霧中

>
>
> David Booth, Ph.D.
> HP Software
> +1 617 629 8881 office  |  dbooth@hp.com
> http://www.hp.com/go/software
>
> Statements made herein represent the views of the author and do not  
> necessarily represent the official views of HP unless explicitly so  
> stated.