Encoded Resources


Several types of Character Entities (Names), are encoded with small, one to three character codes.  That would include, at least, Country Codes (ISO 3166-1), Language Codes (ISO 639-2) and Currency Codes (ISO 4217).  Many of these encoding schemes have been in service for many years, and it only makes sense for Strategic Planning resources to treat the code sets, not as a legacy to be squeezed in, but rather as translation tools which can be used to describe a sensible scope for the plan.


XML inherits automatic Character Entity Processing from Standardized Generalized Markup Language (SGML).  An entity table is a grid with unique value codes for entities of the same class.  Substitution of the code for the entity name in a display appears to be almost instantaneous to the eye.  But the entity table is flat – all codes are unique and peers.  Several XML based formats have been developed to give the entity table depth, SKOS, RDF, OWL, Dublin Core (DCMI), etc..  When used for this purpose, all of the formats are complementary, if implemented properly.


The entity table documents look pretty much the same, a list of the entity values, an XML file to show they look, and and a schema file to relate one subclass of entity to another.  


For the Country Codes, the meta data “header” section of the schema will look something like this:



The text (RDF, with DC+SKOS) which generated this graph (thanks W3C) is at the end.  It spite of the complex look, the header is actually pretty standard fare for the meta data a document should have.

It has the usual who, what, when, where, why details (to the left), a list of key words for “discovery” (in the center right) and a table of contents /classes/concepts (to the right).


With MARC, MODS, The Journal Publishing Suite, etc., it is doubtful that the US Government needs another header section format.  However, this header is based on Dublin Core and SKOS and should be readily convertible to the more specialized formats.  There will not be a great number of these code sets.  For a generalized web interface, embedding MODS in browser friendly HTML in the same way as MathML and SVG, is probably a better way to go.