Structure the data, the how is secondary

Lets define an abstract Example Structure with three elements:

• date
• creator
• version

In the examples, the data in these elements are:

• date=2011-08-11
• creator=M.T. Carrasco Benitez
• version=1

The Example Structure can be implemented in several formats:

• key-value pair
  • in one file
  • file-content pair - the key is the filename, the value the file content
    • date.txt
    • creator.txt
    • version.txt
• Delimiter-Separated Values (DSV) ¶ - a general case of comma-separated values (CSV) ¶
• HTML - view code
• XML
• SQLite ¶ - script to create the database db.sqlite and instructions for viewing the content

The Example Structure is just a 3-tuple; in database terminology, a record (row) with three fields. Most of the above formats are an overkill; for example, DSV is really more appropriated for tables with many records. XML is even more an overkill as it is designed for more complex trees. It is trivial to transform among these formats.

The file-content pair feels rights (and sits on the floor) with the web philosophy: just a URI gets the data directly (no additional programing is needed); for example, give me the version.