Re: TAG Issue proposal: URIs should not be hierarchical

>Do you think about names in terms words written with letters? or
about concepts?

Names are simply labels (think of sticking a paper label to a conserved food
jar) for concepts (be they physical or abstract, general or particular).
Other way of saying just about the same, is that they are symbols.
And yes, most normally, they are arbitrary. Meaning is just by social
agreement.
And yes, they can be encoded into sounds, into shapes called letters, into
whatever.
When thinking "names", the most common case is physical, particular things,
as in "Your cat's name is Felix".

>I'm trying to figure out why you think that names are or are not
hierarchical.

Imagine someone in a kitchen yelling:

  "Hey, you, /animal/cat/felix , get outta that fish!"

If that's the way you'd say it, then names are hierarchical.
If you would just say "Hey, you, Felix, get outta that fish!" , then they
aren't.

>There are structures in things which  help us to understand them

I'm totally aware about things like prefixes and suffixes, composition
rules, etc.
In fact, I have a lot of ideas (and the most solid deployment strategy I
know of) for an artificial world language.

>> + Same as in XML, adding *one* organization means either to leave out all
>> other possible ones, or creating a lot of URI's for the same
>> resource, one for each possible organization. Is this desirable?
>> + I suggest that other than the word "locator" in "URL", the other
>> main reason people think of URI's as locations instead of names, is the
>> existence of paths (regardless of wether they are correct in doing so or
not).
>> + With the advent of the semantic web, we can have all
>> organizations and relationships via metadata. No need to push one (and
only one) >> into URIs (or to create tons of URIs for the same resource).

> I think your issue in the understanding is about mixing

> mechanism to create unique identifiers (meaningless)
> and
> giving meaning to these identifiers.

I'm trying to separate those two things. Names should be just arbitrary
labels (ideally unique identifiers), period. I.e, "felix".
No attempt should be made to plug a part of the ontology into URIs to try to
give them meaning. I.e, "/animal/cat/felix".
The place to do that is the ontology, not the label (the URI, the name).
With the advent of the semantic web, we can acess the node with a given
(unique) name, and then navigate around the ontology from that node to our
hearts content.
Fron node "felix", via "isinstanceof" we *can* get to cat and to animal, no
problems.
No need to plug that in the URI.
But crucially, we can follow ANY trajectory around the ontology we might
want. NOT just one. Not just /animal/cat/felix , but also /furry/black/felix
, multicellular/felix , etc etc.

>> But is it really "convenient"?
>> Names are not trees. And we do not represent ontologies via trees,  but
via graphs.
>> Graphs can be navigated any way the user prefers, not just one given by a
webmaster.

yes it is :) Having hierarchies for writing things are everywhere,
ordering. If you have a *personal* bookshelf, right now near you,
look at it. You made your own classification, your own hierarchy. It
doesn't have a meaning for the rest of us, but it's practical for you.

But precisely. In a tree, we can have only ONE classification. In a graph,
we can have MANY different ones.
To try to plug one classification into the name (the URI) limits us to just
one. It simply is not the place, nor convenient.

I hope this clarifies

Fernando Franco

Received on Monday, 28 August 2006 10:32:07 UTC