------------------------------------------------------------------------
Date:         Thu, 12 May 2005 10:41:26 +0100
From: Andy Powell <a.powell@UKOLN.AC.UK>
Subject: Vocabulary Encoding Scheme vs Syntax Encoding Scheme - summary
To: DC-USAGE@JISCMAIL.AC.UK
------------------------------------------------------------------------

Summing up previous discussion, the test for whether a term is a syntax
encoding scheme or a vocabulary encoding scheme is as follows:

- if the term defines a set of strings, either by enumerating them or by
   providing a rule for building them, the the term is a syntax encoding
   scheme

- if the term defines a class of values (not strings) , then the term is a
   vocabulary encoding scheme.

Therefore, we get the following:

Box             Syntax
DCMIType        Vocabulary
DDC             Vocabulary
IMT             Vocabulary
ISO3166         Vocabulary
ISO639-2        Vocabulary
LCC             Vocabulary
LCSH            Vocabulary
MESH            Vocabulary
Period          Syntax
Point           Syntax
RFC1766         Syntax
RFC3066         Syntax
TGN             Vocabulary
UDC             Vocabulary
URI             Syntax
W3CDTF          Syntax

Note that I've previously argued that RFC1766 and RFC3066 are vocabulary
encoding schemes, but the RFCs specifically state that they define a set
of 'tags', which I can't interpret in any other way than as an enumerated
list of stings - therefore they are both syntax encoding schemes according
to the above rule.

The key difference between syntax and vocabulary encoding schemes is their
place in the abstract model.  Syntax encoding schemes relate to 'value
strings' - vocabulary encoding schemes relate the the 'value' (resource).