Re: Choose your namespace (Was : Personal view)

At 8:34 AM -0700 6/20/00, Henrik Frystyk Nielsen wrote:
>  > But why do this when the binary decision is what we actually _need_,
>>  to satisfy the goals of the namespace rec., and when literal
>>  comparison of absolute URIs enables this to be something that we can
>>  get?
>
>Because you can never guarantee this in a decentralized system. As I
>pointed out, this has *nothing* to do with URIs but everything to do with
>using decentralized names that support indirection. URIs allow up to code
>these names up which no other syntax does.

The current namespace recommendation does guarantee it, because it 
says that you _may not_ use any other form of URI normalization in 
comparison. This is very clearly spelled out, and for clear and 
evident reasons, which have been re-iterated endlessly in this 
discussion.

Literal comparison is _the canonical test_ for namespace identity. 
Not necessarily for the resource denoted by a URI used as a namespace 
identifier. This bothers you, but does not bother everyone.

This is what the namespace recommendation says currently.

If we are forced to accept scheme-dependent comparison as an option 
for namespace processing, the only sane thing to do is to abandon the 
use of URIs for namespaces altogether. However, I don't see a lot of 
support for making namespace comparison variable in this way.

>  > The farther we move from a unique string approach the more kinds of
>>  failures are possible, and the less we meet the _very simple_
>>  requirements for namespaces. A "forbid" solution has none of these
>>  errors, but kills old documents. A "literal" solution is confusing if
>>  you expect absolutization, but is also consistent with respect to
>>  identity.
>
>What you are really saying here is that we should enforce a single URI
>space, call it "ns:" which has certain properties like not use relative
>names and not support indirection. As I have mentioned before, this is a
>GUID. That is a valid thought but please keep it separate from whether URI
>syntax is fine or not.
>
>It's all about choosing your URI space.

No, it's about choosing your comparison algorithm. I don't care what 
results from the de-referencing of a URI given as a namespace. As a 
namespace creator, all I need to know is that I have the authority to 
assign the names that I assign, and that I won't assign the same name 
to another namespace.

But all this philosophy is pretty irrelevant. What matters is that 
namespace names have a single, deterministic comparison algorithm 
that will not change (new schemes being created _cannot_ affect the 
definition of whether two namespaces are guaranteed to be the same).

It's acceptable (if pointless) to encumber the namespace spec with 
rules about the http: scheme (in particular) if that is deemed 
important. However, it's not acceptable if I have the freedom to 
create a new namespace dgd-names:, give it weird equivalence rules, 
and then demand that namespace processors honor those rules. 
Namespaces require a uniform algorithm that will specify when two 
namespaces are equivalent, and when they are inequivalent and that 
_can't_ depend on the URI scheme.

   -- David
-- 
_________________________________________
David Durand              dgd@cs.bu.edu  \  david@dynamicDiagrams.com
http://cs-people.bu.edu//dgd/             \  Chief Technical Officer
     Graduate Student no more!              \  Dynamic Diagrams
--------------------------------------------\  http://www.dynamicDiagrams.com/
                                              \__________________________

Received on Tuesday, 20 June 2000 12:56:42 UTC