- From: Nathan <nathan@webr3.org>
- Date: Tue, 19 Apr 2011 09:42:40 +0100
- To: Lee Feigenbaum <lee@thefigtrees.net>
- CC: RDF Working Group WG <public-rdf-wg@w3.org>, Richard Cyganiak <richard@cyganiak.de>, Ivan Herman <ivan@w3.org>, Dan Brickley <danbri@danbri.org>, Pat Hayes <phayes@ihmc.us>, David Wood <david.wood@talis.com>, Tim Berners-Lee <timbl@w3.org>, Manu Sporny <msporny@digitalbazaar.com>, Sandro Hawke <sandro@w3.org>
Hi Lee, Lee Feigenbaum wrote: > (Commenting against my better judgment...) Appears you have sound judgement to me :) > On 4/18/2011 8:25 PM, Nathan wrote: >> frustrations. Pretty much every member of this WG who I've spoken to, >> read blogs and mails from, or seen commenting on things seems to want >> quite a bit more than we're chartered to do. But, like myself will also >> just stick to the charter and do what is deemed to be needed. > > I am not only content with the scope of work that this group has > accepted, but would be very concerned about the group taking on > additional work. Based on my experience implementing Semantic Web > standards, building software based on those implementations, and working > in W3C working groups, I think making substantial changes to RDF would > be a mistake. Likewise, I agree - changing RDF would be counter productive in many ways - however, I can see areas where a well defined superset would be advantageous (like N3 for instance), and a great deal of benefit to be gained from a strict subset with a clear deployment story (linked data for instance). So, rather than change, perhaps layers of abstraction, information hiding and generalization in order to give less features to more peoples is a more fitting way of putting it. From another angle, there is another form of more generalized RDF which is common to a few specifications and which is required (and implemented) already in tooling and behind the interface which remains unstandardized, the model doesn't handle the many elements required to work with RDF, rather it constrains it to match that semi-practical constrained model which makes it out over the wire in serializations. From another angle, a practical set of limitations such as ground graphs has been adopted by many and used in linked data scenarios, again that remains unstandardized. RDF, as it stands, sits somewhere in the middle, and I believe there's scope to define the aforementioned, any many benefits to be gained by doing that and giving an advanced superset, and a constrained subset, which can then be built on top of, whilst remaining RDF compatible in the common case. > Of course, I've said more than once that it's not my goal to "bring > Semantic Web to the masses." I personally think that while that's a > noble goal, it's not a realistic goal and it's definitely not a > necessary goal for Semantic Web to be (very) successful. And if it were > a goal of mine, I do not think that standards work would be a primary > way to do it. (I'd prefer evangelism, building and promoting tools, and > education.) I totally respect that, and one does not preclude the other :) Bringing the semantic web to the masses may not be the best way to term it, perhaps it's more a case of mixing the following: a - bringing some of the core benefits of the semantic web to the developer masses (shared schemas/properties, uris as names, uniform data model) b - finding the simplest possible common ground which would allow semantic web tooling/people to use the uniform data from (a) c - layering, simplifying and abstracting. I also agree that standards work wouldn't be a primary way to do it, and strongly feel getting working code out there and used is far more beneficial, however it simply would require specification, documentation, review and community input, which is really very close to standardization and follows much the same process. (Also a good portion of it would class as standardization of common practise w/ consideration for practical limitations). > There are a lot of concrete work areas that you identify in your email: > work on them! :) will do > You can even work in them in the context of the W3C, > either via SWIG (mailing list + IRC) or perhaps even via an XG. Yes, most definitely will - there's also the community group approach, really all that would be needed is a mailing list where it won't get in the way or upset the RDF apple-cart, the common project tools, and a little bit of momentum. > Work > with the other people who share your vision to define and implement the > pieces that seem to be missing. Or to catalog existing implementations > that do similar things in different ways and would benefit from a > standard. But, to me, all this is clearly work that should not be done > in the RDF Working Group. Sound advice, the only distinction I'd make is that "should not be done in the RDF Working Group" is different to "of no concern to the RDF Working Group". Hence why I raised here. > I believe that the Semantic Web is a powerful technology space _because_ > different people have very different uses and visions for the > technologies. That's a good thing. But it also means that likeminded > people need to work together to find common ground and make progress, > and those first steps are not well-suited for a standards group. Fully agree :) Thanks for taking the time to reply, positively, Lee. Best, Nathan
Received on Tuesday, 19 April 2011 08:50:07 UTC