W3C home > Mailing lists > Public > public-ldp-wg@w3.org > April 2013

Re: Proposal to close ISSUE-59 (recursive-delete): Reconsider usage of Aggregate/Composite construct to get predictable container delete behavior

From: Arnaud Le Hors <lehors@us.ibm.com>
Date: Fri, 19 Apr 2013 15:14:58 -0700
To: public-ldp-wg@w3.org
Message-ID: <OF1D66558E.38BA3818-ON88257B52.0078DFE5-88257B52.007A3804@us.ibm.com>
Please, let's not rehash the same debate over and over.

>From what I understand we have two opposing views of the world. 

On one side we have people who think that all links are created equal and 
the client/server interaction is solely dependent on the semantic 
associated with the predicates used to link two resources. The client will 
know from the associated semantic whether a link is worth following or 
not, and what following that link means.

On the other side we have people who think not all links are the same and 
the protocol needs to differentiate between those that primarily serve as 
identifiers and those that are there as navigational links. The idea being 
that while the former can be followed for informational purpose they have 
no purpose in the client/server interaction. In this model client need a 
way to distinguish between the two types of links so that they don't get 
lost/sidetracked following links that are not part of the 
protocol/navigation model.

I fear that putting labels on these groups of people will set people off 
but I would say that the RDF/Linked Data (e.g., Henry, Kingsley) fall in 
the first category while the REST people (e.g., Erik) fall in the second 
one. I don't know whether there is a way to reconcile these two views or 
not but I'm sure hammering the same arguments over and over isn't going to 
make a difference other than wasting more bandwidth and mailbox storage.

So unless someone has a suggestion on how to converge, please, let's put 
this topic to rest.

Thanks.
--
Arnaud  Le Hors - Software Standards Architect - IBM Software Group


Kingsley Idehen <kidehen@openlinksw.com> wrote on 04/19/2013 02:39:54 PM:

> From: Kingsley Idehen <kidehen@openlinksw.com>
> To: public-ldp-wg@w3.org, 
> Date: 04/19/2013 02:40 PM
> Subject: Re: Proposal to close ISSUE-59 (recursive-delete): 
> Reconsider usage   of Aggregate/Composite construct to get 
> predictable container delete behavior
> 
> On 4/19/13 5:21 PM, Wilde, Erik wrote:
> > hello kinsgley.
> >
> > On 2013-04-19 12:23 , "Kingsley Idehen" <kidehen@openlinksw.com> 
wrote:
> >>     A directory has a *containment* relation with resources.
> >>     A directory also has symbolic link (so to speak) relation with
> >>     resources.
> >>     The relations above are inescapable. To ignore them is to 
introduce
> >>     regression or undue limitations. Henry explained the logic with
> >>     clarity.
> > it all depends on use cases, at least for the REST side of things. if
> > you're just browsing, you don't care about the difference between a 
file,
> > a hard link, and a soft link.
> 
> Correct.
> 
> >   once you start doing other things (such as
> > being able to add something as content, hard link, or soft link; or
> > removing something and understanding whether you're just unlinking or
> > actually deleting it), that difference becomes relevant and has to be
> > exposed.
> 
> Yes, and it is exposed by the semantics of the relation, as outlined in 
> the ontology by Henry.
> 
> 
> > typically in REST, you start with use cases, design stateless
> > hypermedia-based application flows, make sure you have the interaction
> > data model and interaction affordances in place, and then you have a
> > working design (for that use case).
> 
> Yes, and with RDF based Linked Data there is a fundamental affordance 
> called: machine-readable and comprehensible relation semantics. The 
> schema is logic i.e., First-order logic. The aformentioned affordance is 

> intrinsic to RDF based Linked Data.
> 
> You can look at it this way:
> 
> RDF based Linked Data delivers an affordance that enables URLs denote 
> Web documents that describe entities (which may or may not be Web 
> accessible) via URIs i.e, said URIs are denoted by URIs. If you don't 
> agree with this affordance then you simply have to toss RDF based Linked 

> Data  out of the window re. this entire effort.  To call such an effort 
> the "Linked Data Platform" and then also claim it is based on RDF,  and 
> in line with TimBL's Linked Data meme etc.., becomes a wasteful 
> distortion or reality.
> 
> A Generic Linked Platform is not an RDF based Linked Data Platform. The 
> affordances aren't isomorphic.
> 
> I suspect Henry and others are exhausted, hence their silence re. this 
> permathread.
> >
> > cheers,
> >
> > dret.
> >
> >
> >
> 
> 
> -- 
> 
> Regards,
> 
> Kingsley Idehen 
> Founder & CEO
> OpenLink Software
> Company Web: http://www.openlinksw.com
> Personal Weblog: http://www.openlinksw.com/blog/~kidehen
> Twitter/Identi.ca handle: @kidehen
> Google+ Profile: https://plus.google.com/112399767740508618350/about
> LinkedIn Profile: http://www.linkedin.com/in/kidehen
> 
> 
> 
> 
> 
Received on Friday, 19 April 2013 22:16:18 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:11:46 UTC