W3C home > Mailing lists > Public > public-lod@w3.org > June 2013

Re: Linked Data discussions require better communication

From: Hugh Glaser <hg@ecs.soton.ac.uk>
Date: Fri, 21 Jun 2013 18:41:36 +0000
To: Kingsley Idehen <kidehen@openlinksw.com>
CC: "<public-lod@w3.org>" <public-lod@w3.org>
Message-ID: <49470778-A007-4DAD-BCDD-9CD468B8CBC8@soton.ac.uk>
Hi Kingsley,
Since you attempt to quote/paraphrase my words, I think I need to respond.
In general, I try to avoid repeating myself on the list - I think if I can't communicate what I want in an exchange of perhaps 5 messages, then 50+ is not going to do any better.

David's message, I think, succintly captures some of the dissonance going on for some people:- a view that Linked Data came out of the Semantic Web, and is in some ways a specialisation of the Semantic Web. So it inherits the RDF property from the Semantic Web. So, for people who have such a view, many of the arguments here are equivalent to saying that the Semantic Web is not based around RDF.

Of course the preceding paragraph has some interesting issues in it - all I mean to do by it is to comment on some root of the dissonance.

Anyway, to get to your reference to me, I didn't say that "folks on the Semantic Web list didn't want to discuss" Linked Data.
For the record, this is what I said:
"In fact, when I venture onto the SemWeb list I try hard to avoid Linked Data assumptions, since my understanding is that many members do not find the Linked Data stuff directly interesting, and the focus of that list is best kept separate from Linked Data, in particular with respect to resolvable URIs."
It is sort of disturbing that this statement of mine was posted in a message pointing out that you were misrepresenting my views in an earlier message; this was a misrepresentation that David Booth also commented on.

By the way, Kingsley, your response does not seem to me to address David's message, but seems more to simply repeat your generic statements about your view of the world - a view that I think I understand quite well because I have seen it many times before.

While I am venturing into the discussion, In case people had not realised, I will point out that your references are very dubious.
> 1. http://www.slideshare.net/cloudofdata/why-linked-data-2457996/12 -- "the Web done right" quote by TimBL circa. 2008
> 2. http://www.slideshare.net/derivadow/linking-bbccouk-to-the-linked-data-cloud-1223984/2 -- ditto
> 3. http://www.slideshare.net/danja/linked-data/4 -- "the Semantic Web done right, and the Web done right"  quote by TimBL circa.

At first sight these seem to be quotes from TimBL - after all they are documents with quotation marks and Tim's name on them.
However, on examination, the references are to documents written by people who are attempting to quote or paraphrase something they may have heard from Tim.
It seems to me that adding such references lends a spurious weight to any posting - this message is in response to exactly the sort of errors that can occur when people try to state others' views, in this case you mine.

As I have said before (sorry to repeat myself), I am not really interested in the history for answering the question that interests me.
But in relation to the 5 stars debate, your third reference is quite interesting from the point of view of what Tim thought they meant:
> 4. http://www.youtube.com/watch?v=6YKPqx3FXE4 -- TimBL describing Linked Data

"Linked Data is just the highest quality data you can get from the point of view of people reusing it. If you just put it out there in any old format, then typically someone will have to convert it into some <inaudible> format before they can connect it to other data."
If I cared (I don't really at the moment), I could only interpret this, especially with his use of the word "highest" to mean having to have all 5 stars to be Linked Open Data.
It would be pretty strange to state that 1 star was the "highest quality data", but 2-5 stars was, er, better than "highest quality data".

Cheers
Hugh


On 21 Jun 2013, at 13:01, Kingsley Idehen <kidehen@openlinksw.com>
 wrote:

> On 6/20/13 10:40 PM, David Booth wrote:
>> On 06/20/2013 04:46 PM, Melvin Carvalho wrote:
>>> It's taken 5 years+ to simplify things down to the
>>> Linked Data meme ("The semantic web done right") and is starting to gain
>>> some traction.
>>> [ . . . ]
>>> The beauty of of LD is that it's simple and can be understood by a wide
>>> range or people (especially outside academia).  In terms of branding
>>> it's valid to feel that conflating LD and RDF would be a premature
>>> optimization.
>> 
>> But if you *believe* that Linked Data is "the semantic web done right" then unless you intend to re-architect the semantic web, RDF is *essential*, because RDF is the universal data model that was chosen for the semantic web.  And without a universal data model you get walled gardens of data that cannot be used together.  And that is the *opposite* of what the semantic web is all about.
>> 
>> This point was more fully explained here:
>> http://lists.w3.org/Archives/Public/semantic-web/2013Jun/0120.html
>> 
>> David
>> 
>> 
>> 
> David,
> 
> The problem here is that *essential* is ambiguous in this context. I say that because it implies SHOULD to some and MUST to others. It's much simpler, and less problematic, if one uses less ambiguous words, in this kind of context.
> 
> Linked Data isn't necessarily about the Semantic Web, but that doesn't dissociate it from the Semantic Web. As I've said in the past, the Web isn't necessarily devoid of entity relationship semantics, the issue is that granularity and perceptibility of said entity relationships on the Web varies.
> 
> If I recall, Hugh Glaser even commented about folks talking about Linked Data on this list because folks on the Semantic Web list didn't want to discuss it (or something along those lines). Pat Hayes made reference to Schema.org and RDF, which I pointed out to him is more RDF and not Linked Data, I provided a simple test to make my point.
> 
> The Web is like a massive jigsaw puzzle game. Each resource is a puzzle-piece. Once we step beyond the boundaries of how its wired (technically and philosophically) we end up with something that doesn't work, due to the technical and social inertia that manifests as a consequence.
> 
> When I look at the Web the following patterns stick out to me as being critical:
> 
> 1. "view source" -- enabled use of HTML without actually knowing HTML en route to World Wide Web reaching critical mass
> 
> 2. file create, save, and share -- for collaborative creation and sharing of documents comprised of structured content (this is still a pursuit around with RWW and LDP related efforts are focused)
> 
> 3. "just do it!" -- a simple mindset whereby folks just get on with stuff without seeking the permission of controllers  (DBpedia and LOD are classic exemplars and reminders).
> 
> 1-3 are extremely crowd-friendly.
> 
> You see, anything that impedes the process of crowd engagement is ultimately bad for the Web. This is why (I believe) all attempts to mandate RDF in the context of Linked Data simply stifle the crowd-friendly (and tolerant) nature of the Web that's critical to any of its core components.
> 
> The Semantic Web and RDF etc.. are labels respectively associated with aspects of the Web's broader vision and a collection of tools for getting there. If everyone is exploiting the Web do we really care about their preferred label?
> 
> Today, we are pretty much close to what many sought in the early days of this journey i.e., unambiguous names for everything which includes concrete and abstract concepts.
> 
> Documents, Hashtags, Relations, Relationships, People, Music, Books etc.. are all denoted using de-referencable URIs that resolve to useful information. The Web has actually tweaked itself by leveraging the dexterity inherent in its own architecture. Those who appeared to be trying to break the Web have ended up upgrading the Web.
> 
> Facebook (5-star Linked Data and new support of hashtags), Google (schema.org, GKG, and hashtags), and Twitter (via hashtags) are examples of Data Spaces on the Web with ever increasing resolution -- with regards to the entity relationships semantics that underlie their Linked Data (resources or data objects).
> 
> The industry behemoths mentioned above are all using a data model that's compatible with RDF i.e., that are all exposing triple (or 3-tuple) based entity relationships that are increasingly accessible via HTTP URIs.
> 
> BTW -- in the early days of the Linked Open Data (LOD) effort, TimBL used to say "this is the Web done right" [1][2] . As you can see, there is a theme and a pattern here  -- once you look at the data that's already on the Web.
> 
> Links:
> 
> 1. http://www.slideshare.net/cloudofdata/why-linked-data-2457996/12 -- "the Web done right" quote by TimBL circa. 2008
> 2. http://www.slideshare.net/derivadow/linking-bbccouk-to-the-linked-data-cloud-1223984/2 -- ditto
> 3. http://www.slideshare.net/danja/linked-data/4 -- "the Semantic Web done right, and the Web done right"  quote by TimBL circa.
> 4. http://www.youtube.com/watch?v=6YKPqx3FXE4 -- TimBL describing Linked Data
> 5. http://www.w3.org/2005/Talks/1110-iswc-tbl/#(7) -- Linked Data explained in a single slide
> 6. http://www.slideshare.net/derivadow/linking-bbccouk-to-the-linked-data-cloud-1223984/3 -- basically, URI everything and everything is cool.
> 
> -- 
> 
> 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, 21 June 2013 18:42:16 UTC

This archive was generated by hypermail 2.3.1 : Friday, 21 June 2013 18:42:16 UTC