RE: [bt-spec] addresses and identification

[Peter Furniss <peter.furniss@choreology.com>]

> I'm going to try to address the differences between the two schemes being
> proposed for URIs and their usage (and no pun intended).
>
> HP would have Identifier+endpoint as a URI as the way in which to uniquely
> identify and contact some entity (how the entity is found is down
> to how the
> information within the URI is interpreted. Multiple additional (and
> optional) URIs are also used to locate the "same" entity.
>
> Choreology would have Identifier to identify (a URI) and address as a
> separate piece of data.

That is an accurate statement of our proposal. By virtue of its use ("remain
globally unique and persistent even when the resource ceases to exist or
becomes unavailable") it is a URN.

> Now it would appear at first sight like the same amount of information is
> being used, but just represented in a different manner. For example (and
> let's forget about the format of the address for now as it
> shouldn't have an
> affect on the principle point in this discussion):
>
> HP scheme) http://myaddress.com?1234 http://myotheraddress.com?1234
>
> Choreology scheme) 1234 http://myaddress.com http://myotheraddress.com
>
> However, there is a subtle difference between the two forms which I think
> you have been overlooking and which we obviously haven't described well
> enough: your scheme assumes that the entity is identified in the
> same way at
> all addresses - this need not be the case.
>
> Take the example of someone (Fred) who can be contacted in
> multiple ways: 2
> phone numbers and 1 mailing address. This individual has a wife and two
> mistresses and for reasons known only to him, he has given a
> different name
> for himself to each. So:
>
> his wife knows him as: phone://123456?Fred
>
> mistress one as: phone://09785?John
>
> and mistress two as: snailmail://pobox75?William
>
> All the necessary information is present in each identity to uniquely
> contact the individual if we use the HP proposal.

This pattern seems to contradict :

PRF: > Questions:
     > are your "optional URIs for addressing" used for identification in
any
     > circumstances ?  (i.e. are they specified as being URNs for the
(same)
     > identified thing as payload-id identifies, as well as URLs)

ML:  No, they are used purely for addressing.

In your proposal, quoting your summary above

   Multiple additional (and optional) URIs are also used to locate the
"same" entity.

is "locate" equivalent to "identify and contact" (as the primary URI does)
or only to "contact" ?


>                 In the Choreology scheme
> this simply wouldn't be possible unless we decide to have a matrix of
> identifiers and their associated addresses - starts to look extremely
> complicated. (If you've ever tried to get mail addresses from yahoo,
> freeserve, aol etc. you'll know that it can be an iterative process to get
> an address you want from one and then you find it's already taken at some
> other site - so you end up having different identities and contact point
> information even though people who know about all of your addresses still
> know they refer to you.)
>
> The alternative is that we agree on some compound format for a
> URI (address
> and location) and work purely in terms of these. Then if the same name is
> used at different locations, or different names are used at different
> locations, it just falls out of the mix naturally.

We deliberately wanted to make the identifiers single, regardless of how you
get there. Otherwise the testing to see if is the same thing becomes
complicated - for specification, for the matching, and for the carrying of
the identifiers in CONTEXTs etc.

Peter

------------------------------------------
Peter Furniss
Technical Director, Choreology Ltd
web: http://www.choreology.com
email:  peter.furniss@choreology.com
phone:  +44 20 7670 1679
direct: +44 20 7670 1783
mobile: 07951 536168
13 Austin Friars, London EC2N 2JX