Raw

Koranteng Ofosu-Amaah discusses the following :

• Identification Of Resources
• Manipulation Of Resources Through Representations
• Self-Descriptive Messages
• Hypermedia As The Engine Of Application State

Fielding's magnum opus was published in 2000 and not only has it stood the test of time, it generally seems to have matured with age, with deployments gradually tending closer towards the patterns it describes.



There is one little thing that does seem to have eroded somewhat (though I can't remember offhand how much weight this is given in the thesis): use of the media (mime) type. Around XML this bit of representation metadata has become incredibly unreliable. I had a good (wine-lubricated) argument about this with Leo Simons when I was in Rome the other week. My pov was that it's still really important, Leo's was more that it's not actually necessary. I came away unconvinced, but considerably less sure of my position.



Going back the elevator list, I've personally never really be very sure about what Hypermedia As The Engine Of Application State actually means for the Web, even less so when dealing with Semantic Web material. "Hypermedia" seems a bit doccy for my datahead tendencies. Whatever, there's a neater elevator pitch in those slides of TimBL's I've been linking to a lot of late:

Web architecture 101

• Things are denoted by URIs.
• Use them to denote things.
• Serve useful information at them.
• Dereference them.

Hmm, "useful information" - that would include data, right..? So if the data contains a liberal sprinkling of dereferenceable URIs (which RDF should) then I guess that makes it hypermedia...@en