Lately I came across TBL's Closed World Machine (cwm) and found it fascinating because I could use something with similar capabilities (but much better performance) in my toolset.
I was particularly fascinated by N3, which isn't just a format for RDF, but also comes with a data model that extends RDF in the direction of first-order logic and in which there are many potentials for expressing rules, reification and so forth. Connected w/ that there's a literature on the use of legacy theorem provers and reasoners on RDF data.
It seems like work on that died out around 2006 and now there's the focus on OWL, Description Logics, SPARQL, SPIN, etc. Many RDF tools claim to support N3 input and output, but they don't support the expanded N3 data model, so that's a bit bogus. Is it fair to say N3 is dead now, or are people still doing things with it?
I wouldn't say that Notation3 is dead. In the Semantic Web community it is widely utilised in conversations (in companion with Turtle). I still utilise N3 and Turtle on a daily base and insist that N3 is the better RDF ;)
PS: Turtle is now at least a W3C working draft. Let's see how long it will take until Notation3 will be a W3C working draft (at least).
As a tool vendor we have just decided to remove "N3" support from TopBraid, because for us it was never more than Turtle anyway - Jena only covers the subset of N3 that is equivalent to Turtle. We found that the mix of N3 and Turtle has been causing confusion, and I am glad that the W3C is now making Turtle an official standard. With the rule portion of N3 I believe it may have some benefits, but on the other hand side it would be counter productive to support too many different rule formats at the same time. As you may guess, our bet is on SPIN which simply uses SPARQL syntax for rules.
Here's a couple of links to some N3 tools that I believe are reasonably complete:
for me is less important that n3 is a spec or not as long i can use it and do what i need with it.
I decided to answer mainly because i use all three tools, cwm and both mentioned by @rob-vesse, eulersharp and eulergui, all based mainly on n3 and in my opinion less known, :)
- cwm is in active development, even if there is not visibile, is updated with last changes in python, and bugs are fixed as far are discovered. Probably will be added few new things in the future but this is statement is based more on some guesses, :)
- eulersharp is a good complement for cwm, is a backward chaining reasoner, can use external plugins writen in prolog yap, that means can be extended without need to change eulersharp, plus is the fastest reasoner i know
- eulergui, it use both cwm, eulersharp, plus drools, it make possible to use chained projects, that means after applying rules from one project you can send output to another and fire execution of a chained project.
i don't use fuxi, but if i'm not wrong it use n3 too so i'm sure n3 is alive and happy, :)
we have the habit to consider that something is alive when we see that is often changed but i'm not sure that this is realy true, :)