that ess in 'https'

Joaquin Miller joaquin at netmesh.us
Tue Jun 27 03:08:13 UTC 2006


For a while here I thought i understood what the topic of discussion 
was.  Now i feel i may be missing the point entirely.

I thought we were trying to decide whether we should adopt the 
convention that pairs of URLs like the following two URLs, when used 
as identity URLs, are equivalent: identify the same persona.

   http://joaquin.net
   https://joaquin.net

But much of the discussion seems to be about identifying a Yadis 
document or a web page.  About whether we can suppose that the URLs 
will return the same document to a GET.  Or about whether they  s h o 
u l d  return the same document.  Or about what to make of it if they 
happen to return different documents.  Or how best to tell if the 
documents returned are equivalent.

Now, i suppose those discussions would be relevant if we were trying 
to meet the reasonable expectations of programmers and architects.

But i thought our audience was
   the people who
      we would like to see
          use URLs to identify their personas.

In that case, most of the discussion does not feel to me like it is on target.



How about let's stick for a day or so to the question: How will our 
audience best be served: by the two URLs being equivalent, or by 
their not being equivalent?

How will our audience best be served?
The people who we would like to see use URLs to identify their personas.

Cordially, Joaquin

================ there should be no need to read on. ======================

Here is what I thought was being discussed (spelled out in excess 
detail, just be be perfectly clear).

We use URLs as personal identifiers.
   (They don't identify a person, exactly, but instead identify a 
persona being used by a person (or persons).)
     ((Such URLs might also identify some entity that is not a 
person, but that does not matter for us in this discussion.))
     ((We don't limit ourselves to URLs; certainly an XRI; perhaps 
any URI; or IRI. That does not matter for us in this discussion.))

An identifier is an unambiguous name, in a given naming context. [X.902 12.2]

In our case, the naming context is established, in part, by the 
internet standards, practices, and installed infrastructure.
One part of this infrastructure is the maintenance of the domain 
namespace by the installed DNS infrastructure.
 >  I'll call this context 'the DNS URL context', 'the usual URL 
context' or just 'the URL context'.

URLs are resource locators.  Every URL locates a resource.
      ((Or is broken.  We could have a lot of fun arguing about the 
concept, broken identifier, but that will not be useful for us in 
this discussion.))

Further, URLs are a kind of URI.  URIs are identifiers.  [RFC 
3986]  Every URL identifies a resource.
      ((Or is broken.))

In the URL context, a URL identifies a resource.

Arguably, in the URL context these two URLs necessarily identify 
different resources:
   http://joaquin.net
   https://joaquin.net
     ((For us, in this discussion, that does not matter at all, one 
way or the other.))

But we also have another context, more interesting.  The context 
provided by Yadis and by the various internet identity services, such 
as OpenID, LID public key based identity services, MyID, Sxip, and so on.
 >> I'll call this context 'the Identity URL context'.

I maintain a spiffy little web page located by the URL 'http://joaquin.net'.
I maintain a spiffy little resource identified, in the URL context, 
by the URL 'http://joaquin.net'.

At the very same time, I use that very same URL, 
'http://joaquin.net', as an identity URL: an identifier of the 
persona i present at that web site.

In the URL context the URL 'http://joaquin.net' identifies my spiffy 
little resource.
In the Identity URL context the URL 'http://joaquin.net' identifies 
my little persona.

This amazing feat is possible because, in the Identity URL context, 
URLs identify personas.
Not resources.

I am free to use a different identifier, 'https://joaquin.net', to 
identify that same little persona.

Whether, in the URL context, 'https://joaquin.net' and 
'http://joaquin.net' do or don't identify the same resource does not 
matter one whit, if i happen to choose to use them to identify, in 
the Identity URL context, the same persona.

And, since we are inventing this space out of whole cloth, we are 
perfectly free to require that they identify the same persona.  It is 
simply a matter of our specifying a rule of the naming context.

Whether, in the URL context, 'https://joaquin.net' and 
'http://joaquin.net' do or don't identify the same resource does not 
matter one whit, if we to choose to specify that they identify the 
same persona.

Which way we decide might make architects and programmers a little 
unhappier or more pleased, might make a little more or less work for 
us, might lead to more or fewer design and implementation errors, 
might require more or less time, money, or thought.

But none of that matters.

What matters    o n l y    is what will work best for the average 
woman in the street or at the terminal.  Or man.  Or other persona.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.danga.com/pipermail/yadis/attachments/20060626/534eec7f/attachment.html


More information about the yadis mailing list