Semantics (was Re: [microformats-discuss] Web 2.0: Abused)

Danny Ayers danny.ayers at
Mon Oct 3 02:49:34 PDT 2005

On one specific point raised in that article:

"Semantic(s)" may be an ill-defined word in general, but within the
field of computers it does have fairly specific meaning - the logical
interpretation of the thing in question. The semantics of a program is
effectively how it behaves on execution; the semantics of markup is
effectivelythe abstract model of the domain of interest (and the
mapping between the syntax and that model). Most markup languages can
be mapped to fairly simple relational models.  (I'm not a logician, so
this is broad brushstrokes)

IMHO, a critical aspect of semantics on the Web is disambiguation, so
if you have a piece of markup and can deterministically interpret it
within a given model, you have semantics. Without the disambiguation,
you lose the determinism and are left with rules of thumb (scraping).
Automation (the thing computers are good at) gives way to guesswork.
Putting a string like "name" in a CSS attribute may be unambiguous in
a local system, but on the Web at large something like the profile URI
is needed to give it global semantics.



More information about the microformats-discuss mailing list