Re: OWL and RDF lists

Hi David,

> Am 30.08.2022 um 18:05 schrieb David Booth <david@dbooth.org>:
> 
> On 8/30/22 11:46, Paula Gearon wrote:
>>    It would also be really nice to understand if there is a fundamental
>>    reason for this limitation in OWL.
>> I believe this is to help distinguish the ABox from the TBox.
>> . . .
>> OWL needs a list structure in order to keep collections closed.
> 
> This sounds to me like more evidence that RDF should have a built-in list type, rather than the makeshift first/next ladder that is currently used.

You keep beating the drum for built-in list objects and I tend to think you are right. I would however like to better understand what that involves and I couldn’t find much in that respect. In fact the only thing I can find is the proposal by Leigh and Woods from the RDF Next Steps WS in 2010 [0] which is rather sparse on detail.

What I can imagine as needed is:

- syntax, i.e. 
    (@L :a, :b, :c)
  annotating Turtle's list syntax with a declaration 
  that this is a list object of some kind.

- model, i.e.
  the list object itself is opaque like an IRI 
  but can be described with an "ordinary" list
  and thereby accessed.
  The description is derived from the list object alone
  and if it doesn’t describe the object correctly
  that constitutes not a logic problem but a program error.

- querying is probably not different from ordinary lists

- types of lists should be configurable, i.e.
  RDF containers bag, seq and alt (@BL/SL/AL :a, :b, :c),
  OWL axiomatic lists (@OWL :a, :b, :c),

- maybe even configurable semantics, i.e.
  Unique Name Assumption lists (@UL :a, :b, :c),
  Closed World lists (@CL :a, :b, :c),
  Referentially Opaque lists (@OL :a, :b, :c)

Does that make some sense?


Thomas Lörtsch


> David Booth
> 


[0] https://www.w3.org/2009/12/rdf-ws/papers/ws14

Received on Wednesday, 31 August 2022 22:01:36 UTC