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

Re: ISSUE-37: short description of model

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Thu, 28 Feb 2013 07:33:16 -0500
Message-ID: <512F4E8C.2050109@openlinksw.com>
To: public-ldp-wg@w3.org
On 2/28/13 5:08 AM, Wilde, Erik wrote:
> hello roger.
>
> On 2013-02-27 23:09 , "Roger Menday" <Roger.Menday@uk.fujitsu.com> wrote:
>> I am kind of wary of using 'links' because it might be interpreted purely
>> as 'actionable links'. I am referring to "what" happens - which is that
>> the shape of the graph is changed - where the shape is defined by the
>> resources and their properties/attributes/links/arcs/etc... The "how" is
>> then the hypermedia controls (including links, forms), etc.
> a link being actionable is almost the definition of hypermedia, is it not?
> so from my point of view, defining links as those ones that drive an
> application is in line with the hypermedia and REST notion of what a link
> is for.

Yes, but that's only one part of the story, when the subject matter is 
Linked Data. To "denote" or "refer to" doesn't imply any action per se.
>
> maybe the problem is the following: when i say "link", i am driven by my
> understanding of link being affordances for interactions, so links are
> *only* those "arcs" that are intended to be followed as part of the
> application flow. it's what a media types defines as links in the
> vocabulary it is defining. if we had hyperRDF, we would have a framework
> for this, but since we don't, we'll have to define ourselves which of the
> "arcs" are links, and which aren't.

They are also powerful affordances for "denotation" .
>
>> In the above I would add, "actionable link *or form*"
>> Or this that implied ... i.e. a form directs some action via a link (??)
> a form is an actionable link, just like any other. it simply has rules
> attached to it (defined by the media type) what a client should do when
> following that link. so i guess you could argue that a form link is a bit
> more complex in its interaction semantics, but that's it.
>
>> http://km.aifb.kit.edu/projects/numbers/web/n42
>> But, I know what you mean, and it is potentially mis-leading in
>> conjunction with the word 'link'  But you can "arc" to it :)
> but that's not the number 42, it's some URI.

The URI denotes the number 42.

In addition to denoting the number 42, it also resolves to machine 
comprehensible content that describes number 42.

The content is machine comprehensible because its comprised of a 
collection or relations.

The relations are comprised of 3-tuples, when the model is RDF.

The semantics of the relations are also described in the same machine 
comprehensible manner via domains and ranges.

RDF is basically machine comprehensible algebra, which is really a super 
duper affordance long sought in the realm of computing.

You don't need a big DBMS engine or any other piece of software to sit 
between you and your data . Have an editor in hand and you can express 
relations that are persisted to documents that may or may not be shared 
on a network e.g., Web, Internet etc..

> for a link to be meaningful
> when taken out of context, the target needs to use a scheme that is
> context-free, so that i can bookmark that link or maybe send the
> identifier to somebody else and tell them "interact with this resource to
> achieve some result." but i think this discussion is not central to the
> other issues.

In RDF a URI isn't out of context. That's the fundamental essence of 
RDF's data model [1].


Links:

1. https://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html 
-- RDF concepts and abstract syntax .

>
> 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 Thursday, 28 February 2013 12:33:39 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 9 May 2013 13:44:29 UTC