Hello Lorenz, thank you for your detailed response. I know Construct query can not be the alternative of Jena rules, may be not better than it either. But SPARQL is W3C recommended and provide disjunction in shape of UNION which Jena rules dont. . Also, we can delete new knowledge inferred using SPARQL which we cant using Jena rules, due to its monotonic nature.
In my previous project, rules created problem for me ( and for some of my mates) when the rules fired and generated new triples but I was unable to delete it again because rules are monotonic. Currently, I am on a project to study how much "inference " we can get via CONSTRUCT query, if we dont use rules or make its use limited . On Fri, Aug 25, 2017 at 5:35 PM, Lorenz Buehmann < [email protected]> wrote: > > > > > how to get inference > > (as Jena rule do) from CONSTRUCT query > You should really start to understand how a rule engine works. (By the > way, Apache Jena uses the RETE engine as core). > And then, let me ask you some questions: > > * how are multiple rules processed? > * what happens if one rule r1 "depends on" another rule r2, i.e. r2 > produces some triples that are used by r1? > * what happens if there are cyclic dependencies? > > Once you answered those questions, think about how a single SPARQL query > can resp. should solve this problem? > > and why the inference of CONSTRUCT > > is better than Jena rules. > How do you come to this conclusion? Or why do you think so? Did somebody > tell you this? You should start with the comparison of the > expressiveness of SPARQL vs. a rule engine - by the way, Apache Jena > rules is not a standardized rule language. > What about recursion in rules? For example, SPARQL 1.0 can be translated > from/to nonrecursive Datalog with negation. > > Again, why do you compare those two Jena mechanisms? What is the > expected outcome? > > > > On Fri, Aug 25, 2017 at 12:52 PM, Lorenz Buehmann < > > [email protected]> wrote: > > > >> You should try to talk to and collaborate with your team members, Javed > >> Khan and Sidrah Shah... > >> > >> Next, you're comparing a query language with a rule engine - those are > >> obviously different things (although you can do some things also in > >> SPARQL that you can do with Jena rules.) > >> > >> Regarding performance, there is no answer. Obviously in Jena both is > >> executed in Java probably using some shared Jena core functionalities. > >> Nevertheless, there is no need for comparison. > >> > >> The question that comes up here in the mailing list - why are you still > >> asking those basic questions? What is the task of the project? I > >> remember that you used Jena rules all the time - now you're think about > >> using SPARQL? Why? Inferencing and querying are totally different > >> things. So why are you thinking about refactoring the whole project? How > >> much time do you need for this academic project? I mean, it sounds like > >> you're working on it for almost one year now? Is this really necessary > >> and useful to spend so much time on it? > >> > >> > >> Lorenz > >> > >> > >> On 24.08.2017 13:25, tina sani wrote: > >>> Andy, I need some points in which SPARQL query can be better than rules > >> to > >>> get the desired result? > >>> > >>> Is it fast (in terms of time) than jena rules to be executed by the > >> engine? > >>> Second, in terms of expressiveness, how SPARQL is better than the > rules? > >>> One point I think is jena rules does not support dis-junction while > >> SPARQL > >>> does (UNION). , > >>> > >>> On Thu, Aug 24, 2017 at 2:07 PM, Andy Seaborne <[email protected]> > wrote: > >>> > >>>> This has been asked before, most recently 2017-08-14. > >>>> > >>>> https://lists.apache.org/thread.html/ae4c1f5dfbacfb9bf4294a4 > >>>> cf2967e0c7050c81b20f80e6c425a9aa8@%3Cusers.jena.apache.org%3E > >>>> > >>>> Please search the archives before asking questions. > >>>> > >>>> Andy > >>>> > >>>> On 24/08/17 08:50, tina sani wrote: > >>>> > >>>>> What are some of advantages of SPARQL over (Jena) rules? I know only > >> that > >>>>> SPARQL is W3C recommendation and jena rules are not. Similarly, Jena > >> rules > >>>>> are monotonic which does not add/remove any data just infer new > >> knowledge. > >>>>> How SPARQL is more expressive than rules? > >>>>> > >>>>> If we get some inference using SPARQL, how it differs (better) than > >>>>> inference using rules. > >>>>> > >>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_source > >>>>> =link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon> > >>>>> Virus-free. > >>>>> www.avast.com > >>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_source > >>>>> =link&utm_campaign=sig-email&utm_content=webmail&utm_term=link> > >>>>> <#m_-5991634784996563873_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2> > >>>>> > >>>>> > >> > >
