xml-dev - Re: Identity

[Lars Marius Garshol <larsga@ifi.uio.no>]

* Miles Sabin
| 
| The trouble with an idea of identity that ties it up with being able
| to tell whether two descriptions pick out the same or different
| things is that it has the unfortunate effect of relativising the
| idea to differing abilities to pick out differences.

This approach can be taken pretty far. For example, there are
philosophical schools (notably the Zen and Buddhist ones) that hold
that a division of the world into discrete 'things' is undesirable and
thus kills the very concept of identity stone dead. And if you start
to look closely at it there is a lot to be said for this way of
thinking.

However, this discussion seems mainly to have been founded on Kant and
his view that we are subjects receiving impressions of the world and
on the basis on those impressions deciding that the world is composed
of objects with identity.

| Lois Lane belives that Superman isn't the same person as Clark
| Kent. Clark Kent presumably believes otherwise. Which one of those
| two beliefs is true? 

Within that fictional universe, two main approaches seem useful:

 a) Settle for some function that can be applied to sensory data to
    infer from a continuous stream of such data the existence and
    state of various discrete chunks that we can then label with
    identifiers.

    In the example one such definition could be the naive one that we
    use in every-day life to identify people we know. Under this
    definition Clark Kent and Superman are the identifiers of two
    distinct people to everyone except the one person who knows that
    these two are in fact one person and that his correct identifier
    is Kal-El.

    This is not a very desirable approach, since all observers will
    not necessarily agree on the identity of each object.

 b) Alternatively, one can postulate that behind the stream of sensory
    data we receive there is in fact a world consisting of discrete
    chunks and that these chunks have as one of their properties
    identifiers that hopefully are unique.

    In the example this would be as if we used social security
    number[1] to identify people. Clark Kent would readily give his,
    but unless you could entice Superman to give his SSN you would
    never be able to discover that these two are in fact (under this
    definition) one and the same person. And in some cases, such as in
    the case of the illegal alien who washes Kent's office, the person
    has no SSN at all.

    And so this approach turns out to be slightly more desirable,
    provided that everyone is willing to play along. (Where everyone
    includes Superman, illegal aliens and the government.) And for
    those who are not willing to play along it seems to me that
    approach (a) can be used as a fallback strategy.

Now, if you substitute user-agent for person, HTTP response for
sensory data and resource for chunk, then as far as I can tell

  (a) is the situation we have on the web today, and 
  (b) is the one Paul Prescod proposed with his HTTP-ID HTTP header.


None of these approaches really settle the question of identity one
way or the other, but rather do as we do in everyday life: postulate
that there is an underlying identity and proceed with every-day life
with that as an article of faith.

The Buddhists would claim that this leads to eternal suffering, and
the Kantians would claim that this is the best possible approach.  Not
being really into the Buddhist way of seeing things I think I'll opt
for the latter approach.

--Lars M.

[1] Or whatever your national equivalent is. Personnummer, in my case.


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo@ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)