I am putting together a catalog (catalogs by nature are just metadata), and so rather than using a database, I am using an RDF store and SPARQL processor to compile the data. I
Graphity PHP version is an object-oriented framework that combines JAX-RS-style RESTful API with Jena-style RDF API, and is based solely on W3C standards. Disclaimer: I am one of the developers.
It's easiest to use with a remote RDF cloud-store like Dydra.
You might want to have a look at the LOD2 stack we're developing. Further background on the phases (from modeling to publishing) can be found in the Linked Data Life Cycles.
You might want to have a look at the Sesame Windows Client (SWC) (disclaimer: I am one of its developers).
SWC is a graphical desktop app (in .Net) for connecting with a Sesame server (or any SPARQL endpoint), but it also offers options to just create a local triplestore on the fly (without the need for a server). IMHO it's a useful tool for "getting started with RDF".
I've given a couple of talks on this subject titled Everyday Tools for the Semantic Web Developer which cover a variety of useful tools some of which I developed and others from the open source community.
I would strongly recommend downloading and installing Top Braid Composer
It has loads of stuff to make Semantic Web development easier and there is a free edition.