Struggling with fragment identifiers

Even though some of you have already put in quite some effort in trying to
help me understand identifiers, I still find myself struggling with part of
it. And thus, I'm back with some more questions/thoughts, hoping some of
you can provide me some more insight.

>From a practical standpoint I think by now I have come to grips with how
@resource, @id and @itemid are meant to work, how they parse and what can
be accomplished with them. No more questions there.

Yet when it comes to deciding when to use a fragment identifier and when to
use a URI relative to the page, I find I'm still having a hard time
deciding when to use # and when not to. Which for now has resulted in me
using # for any identifier, just to make sure I don't leave anything out in
my attempts trying to provide linked data. But somehow I doubt whether
that's the right approach either.

Now I like to believe I'm almost there (let a guy dream a bit will ya) but
this piece of the puzzle just keeps eluding me and I don't want to develop
a custom of always using # because I fail to grasp the full concept (as
opposed to just the technicalities).

And if there are even further considerations I should take into account
when providing linked data by means of embedded RDFa or Microdata, please
don't hold back.   :)

-- 
*Jarno van Driel*
Technical & Semantic SEO Consultant
8 Digits - Digital Marketing Technologies

Tel: +31 652 847 608
Google+: https://plus.google.com/u/0/+JarnovanDriel
Linkedin: linkedin.com/pub/jarno-van-driel/75/470/36a/