Different schemes for naming and
addressing have been set up for different
reasons, and so have different properties.
For some schemes, for example, there
is a protocol which allows you to
get from a name to where the object
is. For others, there is not.
Attempts by the Internet Engineering
Task Force to develop a taxonlmy
of names have lead to endless discusion,
as simple division is difficult.
This is because different people
are more interested in some areas
than others, and because setting
up a common vocabulary can be difficult,
but also because the actual properties
of a name can be quite complex.
Here are some examples of properties,
with discussion of different points
in greater depth.
- Resolvability
- Can you, given a name,
access the object tiow hich it refers?
- Persistence
- Will the name be just
as valid next week? Will it have
the same properties in 500 yerars?
- Unambiguity
- Does a given name refer
to only one object?
- Uniqueness
- Does a given object have
only one such name?
- Readability
- Can a person, looking
at a name, write it down? Rememeber
it?
- Descriptions
- Other guidelines for
humans
- Structure:
- Does the structure of
the name allow Information hiding
between different parts of the system?
- Coverage of current situation
- Can
the name be used with existing objects
and protocols?
- Extensibility
Can the naming scheme be extended
to handle new objects and protocols?
- Consensus
- Is the naming scheme commonly
agreed and in use?
(Se also: URL spec on requirements
. Up to naming and addressing , on
to existing )
Tim BL