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

Re: Proof: Linked Data does not require RDF

From: Dominic Oldman <doint@oldman.me.uk>
Date: Wed, 19 Jun 2013 15:19:10 +0100 (BST)
Message-ID: <1371651550.53836.YahooMailNeo@web87805.mail.ir2.yahoo.com>
To: Hugh Glaser <hg@ecs.soton.ac.uk>, "public-lod@w3.org community" <public-lod@w3.org>

This email from Hugh for me is absolutely key and while this list naturally talks about a range of theoretical issues (and definitions) I feel that it needs to start coming to conclusions about how it gets from the theoretical discussions to practical and sustainable (sustainability and stability are key) applications that serve a wide range of audience types.

When Hugh talks about sharing a particular view I also think about the need to share particular objectives, and a particular vision, and match this with a practical way forward. When Hugh talks about widening of issues I think about how we are ever going to produce practical applications based on linked data principles, that operate over many different and varied datasets, and which are trusted and robust. It might be worth moving the conversation to think about practical use cases and reaching conclusions about what it would actually take to produce the solutions that are desperately needed, not to satisfy the people on this list (who all share an interest in making this work), but all the people who deserve to receive the benefits that linked data groups constantly promise but haven't yet delivered  - but which are achievable. 

How does my sector create useful applications that operate across the extremely diverse and varied datasets that highly individual cultural heritage organisations produce but which together form a body of work that could revolutionise the way that we work, discover, collaborate and disseminate important information about our world and culture? Simply publishing 'linked data' in an random and uncoordinated way is not enough. Many (including subscribers to this list) are attempting to find a practical route forward and are working hard to create and demonstrate practical solutions (through practical end user applications using RDF and robust contextual standards) and, if necessary, will focus on better practical solutions - but based on firm and solid (theoretical integrity is important and the views of people on this list and others are therefore also very important) foundations. We do this also thinking hard about the type of infrastructure and support
 that we would also need to establish. 
The type of focus that Hugh talks about would be very welcome. I want to contribute to a discussion about RDF and the other standards that work with it to reach objectives and benefits within particular domains, but in parallel also talk about how we join (link) these different knowledge domains together - in a more focused and strategic way.



 From: Hugh Glaser <hg@ecs.soton.ac.uk>
To: "public-lod@w3.org community" <public-lod@w3.org> 
Sent: Wednesday, 19 June 2013, 13:09
Subject: Re: Proof: Linked Data does not require RDF

Firstly, having now read the threads, I thank you all for a lot of very interesting and thoughtful words.
Also, as best I can describe what I think, it seems that David Booth has eloquently said much of what I would say.

I'm not sure if there are many people still reading these threads, but, just diving in by responding to Norman's message, I think the answer does matter somewhat (hence I asked it), although not necessarily because of the two strands Norman mentions.

The question I sort of thought I was asking was indeed a social one.
And very personal.
I want to discuss stuff about Linked Data (with capitals) in a forum with other people who share sufficiently similar views of what Linked Data means so that the discussion is productive etc..
To me, if Linked Data does not (at the moment) assume that RDF is involved, whatever that might mean, then the discussion is unlikely to be productive, and can actually be quite destructive. This is because of a constant widening of issues, rather than focussing on the primary topics.
The threads (plus personal emails) give me a sense that there is a majority who would like to avoid taking the discussions outside the RDF world, but that there is a vocal minority who will resist this at all costs.

I actually think that the question we are trying to answer is what is the scope of this list.
In a strong sense, Linked Data is what *we* decide it is; we don't have to worry about history, or anything like that (sunk costs), but what we want it to be now. That is, for the LOD list - we certainly have no control over what others might mean by it, any more than Engineers in the UK have control over the fact that the person who fixes the home appliances is commonly referred to as an engineer, or Xerox have control over people using the term in common speech to mean copy (or Hoover, etc.).
But we can try to organise our community (exceptionally fragile as it is), so that we can have productive discussions around what a core of people want to discuss under the term Linked Data.

So what should I do? - Remember, I said this was personal.

Well, if the vocal minority decide that they cannot choose to narrow the view they have of Linked Data to exclude the more general stuff, so that discussions are focussed around stuff that assumes RDF, then I will obviously withdraw.
That's fine with me, although I think it will be a shame.
There is another list (Semweb) that will be a better sole home.
Of course, an alternative would be to have a new list, on W3C or elsewhere, such as Google groups.
This would be for Linked Data discussion, with a current assumption of RDF.
We could even call it Linked Data, as opposed to Linked Open Data, which would actually more accurately describe what gets discussed, in some sense.

Best to you all

On 19 Jun 2013, at 12:33, Norman Gray <norman@astro.gla.ac.uk>

> Kingsley and all, hello.
> On 2013 Jun 19, at 12:06, Kingsley Idehen wrote:
>> The issues at hand are as follows:
>> 1. Is RDF the only option for producing Linked Data that's 100% compliant with TimBL's original meme?
>> 2. Are RDF and Linked Data tightly or loosely coupled?
> Those are good and clear, but I think a third issue is:
> 3. Do the answers matter?
> There seem to be two strands in this thread (which I think has now spread across multiple lists).  One strand is concerned to devise a precise definition of what Linked Data means, and hence what's included in, and excluded from, the definition  (call this the 'technical strand'); the other is content to see Linked Data as a rather 'softer' or vaguer thing, concerned with rhetoric, exposition or dissemination (call this the 'sociotechnical strand').
>  * For the technical strand, of course the answers matter, because how else can you decide whether something is compliant with TimBL's meme (I'm not sure that memes include conformance clauses, but we can let that pass...!).  Hence discussion of reasoning, logic, expressiveness, 'overtly RDF', your Venn diagram, and so on.
>  * From the point of view of the sociotechnical strand, the answers don't matter ('distinction without a difference'), because these are non-questions, because 'linked data' isn't something that can be complied with or not.  Or, put another way, concluding that something is or is not officially Linked Data doesn't imply anything important.
> I think there's a certain amount of talking past one another in this thread, because arguments in one strand seem muddled or even mischievous when viewed from the other.
> Does this help this thread at all?
> All the best,
> Norman
> -- 
> Norman Gray  :  http://nxg.me.uk
> SUPA School of Physics and Astronomy, University of Glasgow, UK
Received on Wednesday, 19 June 2013 14:19:40 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:16:36 UTC