Terminology -- another try

Here is another attempt at terminology based on ideas suggested
by various people.  I am attempting to get away from philosophy 
and move towards operational definitions.  Among all the ideas
posted so far, I like Koen's best.  I don't agree with him about
"variant" though.  We can eliminate the word, but it would be 
a mistake to eliminate the concept.  It is central to what we 
are talking about.  

Here are some tentative definitions with comments marked by ##

## The definition of entity should be first.  We seem to agree on
## it.  It is easy to define.  Other things should be defined in
## terms of it.

entity
------
The information or data transferred in an HTTP transaction.
An entity consists of metainformation in the form of
Entity-Header fields and content in the form of an Entity-Body, as
described in section 11.


resource
--------

A network data object or service that can be identified by a URI
(section 7.2). A resource comprises all the entities which have been
or could be produced in response to an HTTP request with a given URI.
Numerous factors can affect the specific entity produced by a
resource.

These include: 
   1) Time -- a resource may change over time, 
   2) Headers -- for example, Accept, Accept-Language, or authentication
      headers may cause different entities to be produced,
   3) Source of request -- e.g. distribution of a document may be limited
      by IP address,
   4) Query part of the URI.
   5) Others?


variant
-------

A variant of a resource is a subset of the entities which comprise that
resource.  It is a subset consisting of those entities which have been
or could be produced in response to requests varying over time, but
with constant URI (except variation of the query part), constant
headers and all with the same request source.  

## VARIANT is the term we have had the most difficulty defining.  I
## think that operationally something close to the definition above is
## forced on us.  For caching to work, a "strong entity tag"
## consisting of a "strong change-indicator" and a "variant-id" must
## uniquely determine an entity within a resource (URI).  The
## change-indicator reflects only the time dimension.  Thus variant-id
## must distinguish entities based on all other dimensions.  Hence, if
## a variant-id determines a "variant" then a variant *must* consist of
## the entities obtained by variations in only the time dimension.

simple resource
---------------

A resource with only one variant.

## Formerly called a specific or plain resource


compound resource
---------------

A resource with more than one variant.

## Formerly called a generic resource.







  John Franks Dept of Math. Northwestern University
		john@math.nwu.edu

Received on Tuesday, 30 April 1996 07:12:23 UTC