Re: What can W3C do to help

On Friday 14. November 2014 12.56.48 Phil Archer wrote:
> For others' benefit, Kjetil has long been a member of the W3C Sem Web
> community and was recently named Semantic Web Developer of the year at
> ISWC.

Thank you, Phil!

> Thanks for this - it's good to see that after, what, 10 years? we're
> still in alignment! By which I mean I agree entirely with your sentiment
> here - and am doing all I can to address it.

Excellent!

> It's a year since we merged the Semantic Web and eGovernment Activities
> to form the Data Activity, a move that was motivated in part by the need
> to, as you suggest, bring regular Web Devs and Sem Web closer together.
> Very recently I spent time at TPAC talking to people like Doug Schepers
> and Robin Berjon - i.e. some of the W3T members who work most closely
> with Web developers - and asked them how they would solve this issue.

Yeah! I've been writing around to conference chairs for the past few months saying that they should invite 
Robin, he's exactly guy we need in the community. 

> Some thoughts that are flying around but that have yet to coalesce into
> something coherent:
> 
> - We need APIs that provide access to data sources that are effectively
> behind a curtain so that the developer doesn't need to know or care
> where the data comes from, just that it comes (and that it's a JSON
> object).
> 
> - Thinking in graphs is alien to many people which is why SPARQL is seen
> as hard, so we should think "NoSPARQL." (I contend that SPARQL is easy,
> it's the graph thing that people find difficult which amounts to the
> same thing but, you know, semantics matter round here and we're
> professionally inveterate pedants).

Yup! And that's exactly the kind of points where we need to fight loudly. And that's where I start disagreeing 
with Robin, but that's also exactly why he's so important, we need the thermal radiation from that clash to 
elucidate things. :-) But no Robin, no clash.

My perspective there is that the world is more naturally a graph than a tree, and rather than making the 
world a forrest, which it isn't and therefore a forrest can't be used to solve complex real-world problems. 
So, we need to figure out how a graph can make sense to developers. And that's a research problem.

But not all are research problems, but largely missed opportunities. Like, for example, if you want to add a 
description to your website, there's dct:description, og:description, schema:description, and probably like a 
1000 others... Who should you use? danbri's answer was "use any" which is the ideal answer, but I fear 
that's only feasible for Google. The rest of us could use a 2000 SLOC reasoner that could make that 
answer the right one in real life too. But that's nowhere to be found. 

I think the Semantic Web is about adapting computing to the real world, one piece at a time, not boiling the 
ocean, but also, not the other way around (where programmers define out of scope real world problems 
where they cannot easily see a solution).


> I believe there is a huge potential here with a lot of cool stuff
> perfectly possible using existing standards and frameworks, but it might
> help to put some of it together.

Yeah, it has to be coherent. And it has to be quick install and configure and well supported. And that means 
it has to go into Linux distributions, because otherwise, deployers have to provide security support 
themselves. Unfortunately, the LOD2 project, which was about putting such a stack together stopped short 
of this. With the next release of Debian, Virtuoso 7 is not going in, just 6. 4store is going to be removed. 
CWM is removed. Semantic Mediawiki was already removed in the previous release.

> Hmm... I feel a workshop coming on...
> Maybe in the first half of next year in Washington DC ;-)

Perhaps! :-)

> Thanks again for raising this Kjetil - we're on it, and your support is
> crucial.

I try my best! :-)

Kjetil