I recently realized that most of my experience with RDF has been with tools that load triples into memory and then work with them there, so I’ve decided to get to know the disk-based triplestores out there better: Jena, Joseki, Sesame, AllegroGraph, OpenLink, Mulgara… let me know if I’m missing anything here.

This is consuming just about all of my free time at the computer (of which I have little lately because of some very long hours for the employer), so I’ve had a lot less time to write for the weblog. When I’ve gotten further with this research, though, I’ll have a lot to write about.

7 Comments

By Eric Schoonover on January 26, 2009 11:16 PM

I’d really encourage you to take a look at Intellidimensions Semantic Server product. It runs on top of SQL server (any edition including Express) and there is a 60 day trial version. They also have a free academic license.

http://www.intellidimension.com/

Note: make sure if you are using SQL Server Express that you pull down the advanced version that includes full text indexing.

By Leigh Dodds on January 27, 2009 7:23 AM

Hi Bob,

If you want to add the Talis Platform to your list of services to explore, then just drop me a mail and I’ll get you set up with a developer account.\

By Bobx DuCharme on January 27, 2009 8:14 AM

Thanks Leigh!

Eric, I may try that, since I do have a copy of SQL Server running.

Bob

By Lee Feigenbaum on January 29, 2009 12:27 AM

Bob,

You’re also welcome to try out Open Anzo - http://openanzo.org - but really, most of these stores aren’t (or shouldn’t be) directly comparable. Most of them have their sweet spot(s), whether it be raw speed, clustering/scalability, federation, enterprise features, collaboration, full-on inference, lightweight (e.g. RDFS) inferencing, etc.

Horses for courses, and all that.

Anyway, drop me a line if you’re interested to hear more about my personal take on various stores’ sweet spots, or, even better, I’d love to hear what you think after you’ve played around some. :-)

Lee

By Taylor on January 29, 2009 11:58 AM

Bob,

I’m totally biased, I like Jena…most of all because I don’t have enough time to learn more that one semweb tool ;-), but also because they have some SPARQL heavy hitters and it supports the latest greatest sparql goodness (updates). So to tempt you along I’ve edited this example for you…

http://tinyurl.com/bzuqrk

I’ve been playing with different ways to code to the jena api, this example is jenabean’s “Thing” which uses simple interfaces to simplify asserting new triples. It makes it easy to polymorph into various vocabs, the library comes with just a few, but it’s very easy to create more.

By Martin Brousseau on February 4, 2009 5:37 PM

Hi Bob,
Don’t forget to add BigOWLIM and Virtuoso to your shopping list.
Known to be among the most scalable triple store. BigOWLIM is using the Sesame api.

By Bob DuCharme on February 5, 2009 12:34 PM

Thanks Martin. Virtuoso was on my list from the beginning. By adding an OWL layer to Sesame, BigOWLIM looks very cool, so I will definitely be playing with it.