Gisle Hannemyr wrote:
> To me, at least, it is intuitive to see the property "creator" linking
> the value "Adams, Ansel" and any reasonable manifestation of a work
> created by him (including digital scans and print reproductions of his
> photographs.  And the "type" property should tell me if we are dealing
> with an original photograph, a binary scan or an offset reproduction).

Whereas to me it is intuitive that if someone is listed as the "creator"
of a resource, then they are the person who created that resource.
Systems I design will take that into account. Systems that you design
will work differently. But I figure it'll all come out in the wash.

> This means that when trying to formulate a search looking for an Ansel
> Adams photograph, I would enter his name (and not the name of the guy who
> did the scanning, or the gal created the offset plates) in the field
> designated to hold the name of the entity primarily creating the
> intellectual content of the resource.

And the system would show you the photograph. If the system was designed
with any kind of intelligence, it would either give you the option to
check for derivative works, or just show you the derivative works. After
all, if you're looking for "Mona Lisa", it's unlikely that you want to
actually acquire the real Mona Lisa (otherwise you'd be corresponding
with the gallery about the purchase price and conditions).

When I'm creating metadata about a resource, I'm creating metadata about
the *resource*. Not the underlying work. The metadata for the underlying
work will be referenced by DC.Source. The copyright on the derivative
work will likely belong to the original creator, thus DC.Rights will
likely contain "copyright ..." or at least "reproduced with permission
...".

> Certainly, if I was looking for some manifestation of a photograph by
> Ansel Adams, it would probably _not_ occur to me that I should ignore
> the "creator" property and instead search for records where his name
> was linked to the "rights", "source" or "relation" property.

Now say the digital image of the photograph has been retouched.
Someone's done something with it that you don't like. Who do you talk
to, in order to check if the copy you have is the same as the original
digitisation that they produced?

"I want a copy of the Mona Lisa, dammit! This picture shows her smoking
a reefer! That's not acceptable!" Do I go beat up on Leonardo da Vinci
for his irreverent work?

> Of course I can design the search engine so that when someone instructs
> it to search for manifestations of the work of "Ansel Adams", it can
> be real clever and "know" that this means that sometimes the name "Adams,
> Ansel" appear linked to the "creator" property, and sometimes it is
> instead linked to "rights", "source", "relation".  Maybe even
> "publisher" or "contributor"?

No - just DC.Source. Don't try to make things difficult.

> But unless the search engine had the natural language abilities and
> knowledge of the the life and work of Ansel Adams ...

WTF?

There is no need for natural language processing. There is no need for
contextual knowledge. There is only need for recognition by the system
designer that people are more likely to be after a derivative or
licencable work, than the original item. So include the option "search
for derivative works" as part of your standard search form.

When you get in a car and turn the key to the "START" position, does it
occur to you to specifically order the electric fuel pump to start?
Since most people just want to turn a key, the designers of the car
figured it was easiest to have the key move past the "ON" position to
the "START" position. In the "ON" position the fuel pump starts. The
procedure of starting an engine is therefore transparent to the end
user.

The same should be true for a resource discovery engine. If you search
for a work, you might be given the option to discover derivative works,
too. Or, as in the car metaphor, searching for derivative works might be
a default action. The repository you're searching might be run by an art
gallery that wants to sell you posters, mugs, placemats, napkins, face
towels etc. So of course, they'll make sure to let you know that you can
wipe your face on your favourite work of art every evening.

> Searching through all those tables to avoid missing a manifestation
> of a work originally created by Ansel Adams would  probably end up
> returning ... all those pictures
> _not_ created by Ansel Adams [because they're owned by the Foundation]

The confusion ends at DC.Source. If you want, you can explore further up
the DC.Source tree, or further down it, but that's up to you in your own
system.

Regards
Alex

-- 
Alex Satrapa
tSA Consulting Group Pty Ltd.
Canberra, Australia


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%