Resources and URIs

I see a number of messages indicating participants struggling with URI
generation.

I think we should not worry too much about all URI problems.

A URI serves for addressing.
In my opinion there are two kinds of addressing 'ranges': let's call 'em private and public.

When fooling around in your own address space (your program, your local document) anything will do as long as you can resolve
all your local addressing. Let us call this the 'local process' space.
You may have a Java program manipulating some local model: why bothering about URIs when nobody else can access your memory space?
Or you may generate an RDF document that serves as a message between two processes, why bothering to make the ID's of the
transmitted resources globally unique? Just call them URI1, URI2 and URI3 as long as the receiver can identify them within the message.

It's another matter when your resources get public: the addressing takes place in a global namespace.
That's when it gets serious. This applies to the 'global persistent' space only.
Let us focus on that problem.

Is it still a different problem than creating unique URLs now?

Received on Tuesday, 7 December 1999 13:26:10 UTC