W3C home > Mailing lists > Public > semantic-web@w3.org > June 2020

Re: defining the semantics of lists

From: Holger Knublauch <holger@topquadrant.com>
Date: Thu, 11 Jun 2020 08:22:23 +1000
To: semantic-web@w3.org
Message-ID: <5411ad39-fe71-8a48-ee32-25f29b32d822@topquadrant.com>
FWIW one approach we are using now for lists is based on Reification, 
simply setting an index property as reified value. Compared to rdf:Lists 
this has the benefit that the list members can still be accessed as 
simple S, P, O triples. Converting the values into an array is O(n) 
assuming the indexes are integers from 0..n. Looking up a specific value 
by index requires a join and depends on how reification is implemented. 
I guess all I am saying is that if reification would be solved (and more 
and more triple stores converge to RDF* as de-facto solution) then it 
can become a natural alternative to rdf lists and containers. Turtle 
could be extended to render them naturally.

Holger


On 11/06/2020 08:12, Thomas Passin wrote:
> On 6/10/2020 5:34 PM, Frans Knibbe wrote:
>> I am afraid that for geographical data (that's my background) a 
>> solution with explicit length properties might not work. That's 
>> because the distance between two locations may not be known or can 
>> not unambiguously be determined. The earth's surface is not smooth 
>> and constantly changing, that complicates matters.
>
> Heh, heh!  I chose not to get into whether or not the triangle is to 
> be in Euclidean space or not.  Talk about potential data explosion!  
> And notice that these possibilities don't have anything to do with the 
> list-ness of the thing.
>
> The style of modeling something is not a one-time thing that has only 
> one possible way, at least not for something complicated.  I like to 
> call these modeling patterns "idioms".  One can usually move details 
> out of the model and into the software that interprets it, if the 
> software understands the idiom.  But then you can't share those graphs 
> except with others whose software understands the idiom.
>
> Personally, before any new list constructs get standardized, I would 
> like a lot more modeling (such as my example) to get done. Once a lot 
> of people get solid on that - i.e., have wrestled with it in detail - 
> would it makes sense to me for shortcuts to get evolved to a standard.
>
> TomP
>
Received on Wednesday, 10 June 2020 22:22:39 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:46:04 UTC