RDF Steps Carefully Forward

April 14, 2011

18 months ago, when Ivan Herman and I began to plan a new RDF Working Group, I posted my RDF 2 Wishlist. Some people complained that the Semantic Web was not ready for anything different; it was still getting used to RDF 1. I clarified that “RDF 2” would be backward compatible and not break existing system, just like “HTML 5” isn’t breaking the existing Web. Still, some people prefered the term “RDF 1.1”.

The group just concluded its first face-to-face meeting, and I think it’s now clear we’re just doing maintenance. If we were to do version numbering, it might be called “RDF 1.0.1”. This might just be “RDF Second Edition”. Basically, the changes will be editorial clarifications and bug fixes.

The adventurer in me is disappointed. It’s a bit like opening your birthday present to find nice warm socks, instead of the jet pack you were hoping for.

Of course, this was mostly clear from the workshop poll and the charter, but still, I had my hopes.

The most dramatic change the group is likely to make: advise people to stop using xs:string in RDF. Pretty exciting. And, despite unanimous support from the 14 people who expressed an opinion in the meeting, there has now been some strong pushback from people not at the meeting. So I think that’s a pretty good measure of the size change we can make.

As far as new stuff…. we’ll probably come up with some terminology for talking about graphs, and maybe even a syntax which allows people to express information about graphs and subgraphs. But one could easily view that as just properly providing the functionality that RDF reification was supposed to provide. So, again, it’s just a (rather complicated) bug fix. And yes, making Turtle a REC, but it’s already a de facto standard, so (again) not a big deal.

The group also decided, with a bit of disappointment for some, not to actively push for a JSON serialization that appeals to non-RDF-folks. This was something I was interested in (cf JRON) but I agree there’s too much design work to do in a Working Group like this. The door was left open for the group to take it up again, if the right proposal appears.

So, it’s all good. I’m comfortable with all the decisions the group made in the past two days, and I’m really happy to be working with such a great bunch of people. I also had a nice time visiting Amsterdam and taking long walks along the canals. But, one of these days, I want my jet pack.

5 Responses to “RDF Steps Carefully Forward”

  1. […] RDF Steps Carefully Forward […]

  2. […] Hawke, also in attendance at the Working Group, wrote down his thoughts in a blog post: “18 months ago, when Ivan Herman and I began to plan a new RDF […]

  3. […] Hawke, also in attendance at the Working Group, wrote down his thoughts in a blog post: “18 months ago, when Ivan Herman and I began to plan a new RDF […]

  4. […] is trying to solve the fundamental problems. The new RDF working group is doing just a maintenance work. The standards are set and gain maturity, but they are based on the […]

  5. Alan Mullane Says:

    Apologies for the length of this reply – I didn’t know where else to post – please bear with me. 

    I come from a sw development product delivery background and while our practices may not be leading edge there is one pattern I’ve noticed over the years – most time during our software development efforts has been on learning tools and navigating the various complexities of closed GUI and document systems to get a few thousand lines of code onto a server so our customers can finally (6 months or a year later and several millions of their well earned dollars) get to use the features provided.

    It would be nice if software development could be described with RDF. It would also be nice if semantic web development could be described with RDF. My interest in RDF may be different from the majority out there who want to build a global semantic web and build reasoners to determine the truth behind data. I’m more interested in a local semantic web and how to use RDF to describe info within a local context that is always true and that allows the specific mundane to be removed from everyday life to allow for more effective, creative and innovative things to get done.

    The problem is that even to grasp the subject of the semantic web and to use it in anger to solve real problems is hard at the moment and I suppose that feeds into working groups not being able to push concepts on as fast as they would like – I find it very difficult also to convince other software devs and other experts in general to use RDF.

    I’ve recently been thinking about using open source command line tools to build other open source tools but also to describe their install and run using RDF and how to apply these to adopting semantic web techniques. This brought me to this question about RDF – is there a template that could allow anyone to easily start building and using RDF online to first define local context RDF and also add global context RDF and also apply this to solve real world problems using java (or other) agents. So if I wanted to get photos from my facebook page and share with a friend who doesn’t have a facebook account, I could do it with this online semantic web environment of open source tool (e.g online tool invokes local ‘get my photos’ agent running on my pc which invokes wget command with ‘–spider’ option to get my photos from facebook and then sends link to my friend with ’email’ agent, all hidden behind semantic web online app). I know how to get to my photos on facebook and using SSO I can tell my local ‘get my photos’ agent to do this.

    I am going to attempt to build such an online app, call it software enterprise environment template – sweet) for allowing adoption of semantic web development by anyone interested in removing the mundane and that is built using RDF and that runs using RDF to evolve a set of existing open source tools into a powereful but very simple online tool – note, no machine learning or nlp required here, just some RDF, a few basic agents to handle download, install, rdf add and view and link, maybe option to add other agents based on these and a few well chosen open source apps. 

    If there is something else like this out there to achieve this I’d be very interested in using it. I would want to run this from a browser on my windows pc (yes, i am a lowly windows pc software developer currently as my current work requires this, normally I work on linux) but it could be run from any browser and connected to a cloud environment – here are the tools i would use – cygwin(wget,gunzip,diffutils), java,umlgraph, graphviz,maven(spring mvc framework,jetty,jena,jena-ld,ssl) and processingjs – the last one would dynamically load png files of RDF graph visualisations generated by umlgraph into the current canvas (html5 is a must also) – if I did manage to build this I would be happy to make it open source so more people could use it. 

    I have to mention umlgraph here – this command line tool (btw, for this to work, all open source tools have to be command line and allow for unattended install to allow agents to manage them) can generate a visualisation of a set of skeleton java classes and can be annotated with relationships and associations hierarchically – therefore it can visualise any type of RDF data, I think (checkout umlgraph.org). This would be a key feature for rapid semantic web adoption in my view – the ability to render linked data visually on the fly as you create it or load it – using the tools mentioned above, this should not be difficult to build. 

    Anyway too much specifics for this blog – I guess the bottom line was that I wanted to share this here with you and get your thoughts on if it is something that you would be interested in or if you know some one who would be interested in using this or if it would be helpful for the RDF working group.

Comments are closed.

%d bloggers like this: