W3C home > Mailing lists > Public > semantic-web@w3.org > December 2012

Re: Well Behaved RDF - Taming Blank Nodes, etc.

From: Lee Feigenbaum <lee@thefigtrees.net>
Date: Sat, 15 Dec 2012 22:19:05 -0500
Message-ID: <50CD3DA9.5020603@thefigtrees.net>
To: Pat Hayes <phayes@ihmc.us>
CC: ross.horne@gmail.com, Hugh Glaser <hg@ecs.soton.ac.uk>, David Booth <david@dbooth.org>, semantic-web Web <semantic-web@w3.org>
On 12/15/2012 2:52 AM, Pat Hayes wrote:
> On Dec 14, 2012, at 9:41 AM, Ross Horne wrote:
>> Thank you for the filtering..
>>> On 13 Dec 2012, at 14:41, Lee Feigenbaum <lee@thefigtrees.net> wrote:
>>>> On 12/13/2012 2:00 AM, Pat Hayes wrote:
>>>>> Another example: a picture of some celebrity standing next to a horse. I have a URI for the celebrity, but I don't have and don't need one for the horse: and if I were to invent one for each horse, then I could no longer query for retrieval of a picture of that person with "a horse", but would have to remember the URi for each of the bloody horses. But nobody gives a damn about the particular horse.
>>>> Could you explain this more?
> No, but I could retract it. You are right, the querying works fine with or without the blank node. But the other argument still holds: I don't need, and shouldn't have to invent, a URI for something that does not need to be identified. Particularly if in fact I cannot identify the thing in quesiton. Inventing a URI to "identify" an anonymous horse or an unknown room in a building is simply a mis-use of the very idea of a URI as being a universal identifier.

Is inventing a URI expensive for you? It's not in most 
systems/environments I've worked with. Are you more concerned about the 
costs of creating the URI or of maintaining it over time? Or are there 
cases for which having an identifier for a definitive-yet-unknown entity 
actually causes harm in trying to work with the data? (I believe that 
the cases in which it is difficult to work with blank node data (e.g. 
not being able to easily do successive queries about the data) are well 


(BTW, I'm not advocating for getting rid of blank nodes. I'm perfectly 
happy not using them myself while others use them in the ways they see 
fit. I'm just trying to understand that latter camp a bit better.)

> Pat
>>>> Because I'm picturing just doing:
>>>> SELECT ?photo {
>>>>   ?photo a :Photograph ;
>>>>        :depicts :ThePerson ;
>>>>        :depicts [ a :Horse ] ;
>>>>   .
>>>> }
>>>> ...which works fine whether the horse is represented with a blank node or a URI.
>>>> Lee
>> Yes! That's right. Compelling use cases for blank nodes are covered by
>> SPARQL. RDF triples *without blank nodes* are adequate and
>> understandable by everyone. Let SPARQL do the work.
>> The List/Collection use case is covered by Turtle where, like in any
>> other language, lists are primitive e.g. ( uri1 uri2 uri3 ). Let the
>> primitives be the canonical form and let the bnode encoding be a
>> historical curiosity. It was a dubious decision to encode lists in the
>> first place.
>> This leads to a half page specification for Well-Behaved RDF with
>> little or no ambiguity.
>> Ross
> ------------------------------------------------------------
> IHMC                                     (850)434 8903 or (650)494 3973
> 40 South Alcaniz St.           (850)202 4416   office
> Pensacola                            (850)202 4440   fax
> FL 32502                              (850)291 0667   mobile
> phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Sunday, 16 December 2012 03:19:35 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:42:38 UTC