FaCT++ is open source and certainly faster for many things.
I'm not a huge Kowari fan...no RDFS and it seems a bit of a PITA to get working. Another category that could be interesting is "lightweight, simple" RDF stores. I.e., for small projects where you just want to read in an RDF/XML file process it, and dump something.
Kendall
· 4 years ago
FaCT++ is open source and certainly faster for many things.
Yeah, I knew it was faster for some things, but I somehow missed that it's open source. Cool. I'm still sticking with Pellet, but I guess now it's more of a homer-choice than before.
As for Kowari, well, yes, it stinks in many ways. But then that was kinda my point; an earlier draft had a bit section on how much RDF databases stink. As you know, they pretty much all suck across the board. But at least they do so in interestingly different ways.
There are lots of other interesting categories, including the one you point out -- in that category I'd name Redland, rdflib, and SWI-Prolog as worth a look-see. That case is, for me, anyway, almost entirely dominated by host language, since it's typically in-process with the code yr already writing or have written.
Dave Beckett
· 4 years ago
Clearly I have an interest in this too :) but more interesting would be expansions on why you might choose a particular OWL reasoner as "best". From what I've read and heard so far, they all emphasise different things, do different sub/supersets of different OWL layers so comparison is rather multi-dimensional. Do you really mean to say Pellet is the "most ... proprietary" or did I read that wrong? Pellet's the best RDFS and OWL editor for me. Jena seems a strong peer of Sesame but isn't mentioned at all. Maybe you could say what Java systems are below Pellet, SWOOP and the SPARQL frontend?
Kendall
· 4 years ago
Dave,
Yeah, lotsa OWL reasoners are focusing on different things, which I take to be a mark of an active, but immature market. That's a good thing and it does make comparisons complicated. I rank Pellet as best because it's open source, reasonably fast, has a very clever, engaged development team, is very feature-rich (I meant to say, btw, that it's probably more feature-rich than any proprietary or open source OWL reasoner), and has an engaged growing community of users. I give a lot of weight to the fact that the Pellet development team is engaged in Semantic Web research at a fundamental level, including stuff like e-connections, which I think are the bee's knees and very promising.
You said "Pellet's the best RDFS and OWL editor for me". Did you mean to say "SWOOP"?
I know about Jena, of course, and have some experience with it, both first and second hand. Let's put it this way: if I were building a SemWeb application and had to use Java, and so had to choose between Sesame or Jena, I would choose Sesame 10 times out of 10.
I don't know what you mean about "Java systems below Pellet, SWOOP, and the SPARQL frontend" -- but I should have mentioned yr SPARQL frontend as an alternative to Leigh's, I just couldn't remember the name and got sloppy.
A link here would be very welcome.
Ian Dickinson
· 3 years ago
"choose Sesame 10 times out of 10" because .... ?
Serious question. Not least because we've heard the opposite opinion (people who've had better experiences building their apps with Jena than Sesame), so there are clearly some sets of criteria that indicate either tool for some user applications. If your interest is in helping semantic web developers make appropriate choices of tools then making your criteria clear would be helpful. Being broad, open source and tracking the standards, the only objective criteria you mention, don't differentiate Jena and Sesame. More generally, it just seems a bit lame to make absolute subjective recommendations without at least some supporting evidence.
Ian (declaration: member of the Jena team)
Maggie
· 3 years ago
I'm also interested in why you prefer Sesame, mostly because you also recommend Pellet and talk about SPARQL. I'm new to semantic web work and have been trying to determine whether to try Sesame or Jena for my project. I've heard good things about Sesame, but Pellet works with Jena, and Sesame's stable version doesn't support SPARQL.
Is there a good OWL-reasoning store for Sesame? (Do you have a recommendation?) Is it annoying to have to use SerQL (or is Sesame 2.0 stable enough to get things done?)
Kendall
· 3 years ago
Okay, so I prefer Sesame because it is better designed and implemented. The Sesame Sail architecture is relatively easy to extend; we did so with surprisingly little effort.
And as a recent RDF database shootout attests, Sesame is significantly faster RDF database than Jena. In our tests it's not even close.
Also, when you've been around software and written enough code, you start to just form a sense for what's good. My nose tells me Sesame is better.
As for SPARQL and Pellet, I probably wouldn't make a decision based on temporary advantage. SPARQL is coming to Sesame sooner than later, and I'd be surprised if one couldn't make Pellet and Sesame play nicely. (It depends on what you really want them to do.)
There are two OWL reasoners listed the Sesame site (BOR and OWLIM). I couldn't make any kind of useful recommendation w/out knowing more about yr situation.
As for query languages, Sesame includes RDQL and SeRQL. I don't know how stable the 2.0 stuff is, but I'd be surprised if it's very long before it's useful in production settings.
But I think of this stuff as a temporary (and limited) advantage at best. :>
Matt Williams
· 3 years ago
Interesting article...I'd heard a bit about Sesame, but will go and have a look.
Agree about Kowari, both that it's good, and a REAL PITA to get running (all the docs are out of date wrt packages, so even running the examples breaks....).
Also agree about Pellet, which is great. AFAIK, FACT++ is only OWL-Lite, whereas Pellet is OWL-DL plus cardinality. Others (like RACER) lose out due to the closed source/ 'license your kidneys to use the software' approach.
I'd choose Protege over SWOOP any day...
Just my penny's worth, Matt
Bijan Parsia
· 3 years ago
FaCT++ is actually now all of SHOIN, thus (depending on datatype support) all of OWL-DL. It's ABox support is weak though and it doesn't (yet) have any optimizations for nominals.
Pellet does not (yet) support *qualified* cardinality, alas. Soon we hope. Pellet does support nominals with a lot of optimizations and nice ABox support (including conjunctive query). Not to mention all the debugging features (though some of them are reasoner independent).
Re: Protege over SWOOP...what can I say, when you're wrong you're wrong ;)
Ok, I hate using Protege and and a main force behind Swoop, so I'm biased. If we can get Swoop stable, I think both the functionality and the usability blow Protege out of the water. (For OWL, at least.)
zzpaul
· 2 years ago
Hi, I am a beginner on ontology/semantic web. I have tried using Protege. I found it pretty good. It is quite easy to learn and use and I like the OntViz that displays an ontology graphically.
For SWOOP, I can't find any tutorial on it. It makes me hard to learn. Do you all know where can I find some tutorials on it? Thank you!
Bijan Parsia
· 2 years ago
zzpaul,
You might look at some of the demo demo movies which, while quite old, should give you some idea.
All Comments
I'm not a huge Kowari fan...no RDFS and it seems a bit of a PITA to get working. Another category that could be interesting is "lightweight, simple" RDF stores. I.e., for small projects where you just want to read in an RDF/XML file process it, and dump something.
Yeah, I knew it was faster for some things, but I somehow missed that it's open source. Cool. I'm still sticking with Pellet, but I guess now it's more of a homer-choice than before.
As for Kowari, well, yes, it stinks in many ways. But then that was kinda my point; an earlier draft had a bit section on how much RDF databases stink. As you know, they pretty much all suck across the board. But at least they do so in interestingly different ways.
There are lots of other interesting categories, including the one you point out -- in that category I'd name Redland, rdflib, and SWI-Prolog as worth a look-see. That case is, for me, anyway, almost entirely dominated by host language, since it's typically in-process with the code yr already writing or have written.
Do you really mean to say Pellet is the "most ... proprietary" or did I read that wrong? Pellet's the best RDFS and OWL editor for me.
Jena seems a strong peer of Sesame but isn't mentioned at all.
Maybe you could say what Java systems are below Pellet, SWOOP and the SPARQL frontend?
Yeah, lotsa OWL reasoners are focusing on different things, which I take to be a mark of an active, but immature market. That's a good thing and it does make comparisons complicated. I rank Pellet as best because it's open source, reasonably fast, has a very clever, engaged development team, is very feature-rich (I meant to say, btw, that it's probably more feature-rich than any proprietary or open source OWL reasoner), and has an engaged growing community of users. I give a lot of weight to the fact that the Pellet development team is engaged in Semantic Web research at a fundamental level, including stuff like e-connections, which I think are the bee's knees and very promising.
You said "Pellet's the best RDFS and OWL editor for me". Did you mean to say "SWOOP"?
I know about Jena, of course, and have some experience with it, both first and second hand. Let's put it this way: if I were building a SemWeb application and had to use Java, and so had to choose between Sesame or Jena, I would choose Sesame 10 times out of 10.
I don't know what you mean about "Java systems below Pellet, SWOOP, and the SPARQL frontend" -- but I should have mentioned yr SPARQL frontend as an alternative to Leigh's, I just couldn't remember the name and got sloppy.
A link here would be very welcome.
Serious question. Not least because we've heard the opposite opinion (people who've had better experiences building their apps with Jena than Sesame), so there are clearly some sets of criteria that indicate either tool for some user applications. If your interest is in helping semantic web developers make appropriate choices of tools then making your criteria clear would be helpful. Being broad, open source and tracking the standards, the only objective criteria you mention, don't differentiate Jena and Sesame. More generally, it just seems a bit lame to make absolute subjective recommendations without at least some supporting evidence.
Ian (declaration: member of the Jena team)
Is there a good OWL-reasoning store for Sesame? (Do you have a recommendation?) Is it annoying to have to use SerQL (or is Sesame 2.0 stable enough to get things done?)
And as a recent RDF database shootout attests, Sesame is significantly faster RDF database than Jena. In our tests it's not even close.
Also, when you've been around software and written enough code, you start to just form a sense for what's good. My nose tells me Sesame is better.
As for SPARQL and Pellet, I probably wouldn't make a decision based on temporary advantage. SPARQL is coming to Sesame sooner than later, and I'd be surprised if one couldn't make Pellet and Sesame play nicely. (It depends on what you really want them to do.)
There are two OWL reasoners listed the Sesame site (BOR and OWLIM). I couldn't make any kind of useful recommendation w/out knowing more about yr situation.
As for query languages, Sesame includes RDQL and SeRQL. I don't know how stable the 2.0 stuff is, but I'd be surprised if it's very long before it's useful in production settings.
But I think of this stuff as a temporary (and limited) advantage at best. :>
Agree about Kowari, both that it's good, and a REAL PITA to get running (all the docs are out of date wrt packages, so even running the examples breaks....).
Also agree about Pellet, which is great. AFAIK, FACT++ is only OWL-Lite, whereas Pellet is OWL-DL plus cardinality. Others (like RACER) lose out due to the closed source/ 'license your kidneys to use the software' approach.
I'd choose Protege over SWOOP any day...
Just my penny's worth,
Matt
Pellet does not (yet) support *qualified* cardinality, alas. Soon we hope. Pellet does support nominals with a lot of optimizations and nice ABox support (including conjunctive query). Not to mention all the debugging features (though some of them are reasoner independent).
Re: Protege over SWOOP...what can I say, when you're wrong you're wrong ;)
Ok, I hate using Protege and and a main force behind Swoop, so I'm biased. If we can get Swoop stable, I think both the functionality and the usability blow Protege out of the water. (For OWL, at least.)
For SWOOP, I can't find any tutorial on it. It makes me hard to learn. Do you all know where can I find some tutorials on it? Thank you!
You might look at some of the demo demo movies which, while quite old, should give you some idea.
Cheers,
Bijan.