User bastian spanneberg - Semantic Overflow

Answer by Bastian Spanneberg for OOP mapping / ORM solutions for Java

2010-07-28T08:59:09Z

If you're a Jena developer, jenabean might be worth a look. It works against Jenas Model API, so it should be easy later on to plug in one of the two jena backends (SDB,TDB) or use some wrapper to interact with another RDF store (SAIL,AllegroGraph).

Answer by Bastian Spanneberg for Sort by date optimizations on large dataset

2010-07-08T14:19:18Z

I discovered a thread on the sparql-dev-mailinglist which also discusses this topic in some detail. It's worth a read, and to take one essential part out of it:

In short: there are a number of interactions between SPARQL and RDF, and design decisions in SPARQL itself, which make it tiresome or impossible as an implementor to make these queries fast. As an industry we simply haven't had the time to push the specs towards efficiency, to learn tricks, or fine-tune.

Answer by Bastian Spanneberg for OWL Full and reasoning

2010-07-08T09:03:55Z

From the OWL Web Ontology Language Guide

OWL Full is meant for users who want maximum expressiveness and the syntactic freedom of RDF with no computational guarantees. For example, in OWL Full a class can be treated simultaneously as a collection of individuals and as an individual in its own right. Another significant difference from OWL DL is that a owl:DatatypeProperty can be marked as an owl:InverseFunctionalProperty. OWL Full allows an ontology to augment the meaning of the pre-defined (RDF or OWL) vocabulary. It is unlikely that any reasoning software will be able to support every feature of OWL Full.

The problem with OWL Full is, that reasoning algorithms cannot make any guarantees about termination due to the above mentioned facts, and that is why most reasoners do not support OWL Full with all of its features.

Answer by Bastian Spanneberg for How do I build a reasoner?

2009-11-04T13:07:20Z

I would start to check out the ALC description logic, which is one of the most simple DLs, and then try to implement some simple reasoning algorithms for it, like the one described in the "EXPTIME tableaux for ALC" paper by Giacomo et al., to gain a feeling and deeper understanding for the matter.

Edit: In the meantime, there was another question on Tableaux algorithms which might also be of interest for people trying to develop a reasoner on their own.

Answer by Bastian Spanneberg for SPARQL DBPedia - Find all persons who were born on this day

2010-07-04T21:32:47Z

Here is a working version which uses only standard compliant SPARQL1.0 without Virtuosos built-in functions. I assume that due to internal optimizations, the built-in functions should result in a faster query execution, however this query works, no matter if Virtuoso is the underlying store or not.

The first filter can also be skipped, assuming that ont:birthDate only points to literals of type xsd:date or xsd:dateTime (which should be te case).

PREFIX ont: <http://dbpedia.org/ontology/> 
PREFIX foaf: <http://xmlns.com/foaf/0.1/> 
PREFIX xsd:    <http://www.w3.org/2001/XMLSchema#> 
SELECT ?name ?date WHERE { 
  ?person ont:birthDate ?date; 
    foaf:name ?name . 
  FILTER( 
    ( ( datatype(?date) = xsd:date ) || ( datatype(?date) = xsd:dateTime ) ) && 
    ( ?date <= "2010-07-03"^^xsd:dateTime ) && 
    ( regex(str(?date), "[0-9]{4}-07-03") ) 
  ) 
}
LIMIT 10

Answer by Bastian Spanneberg for Visualisation toolkits for RDF?

2010-06-25T09:50:35Z

It's more of a quick'n'dirty way to visualize RDF, but you can use Raptor to produce a GraphViz DOT file from RDF which you can then visualize with GraphViz:

shell> rapper -o dot <rdf_input> | dot -Tpng -o output.png

Answer by Bastian Spanneberg for SPARQL for the piece of RDF with same elements

2010-06-21T13:57:19Z

Picking up the comment from castagna, you should think in triples instead of RDF/XML, resulting in a dataset like:

<some_uri> rdf:type <resource_uri> ;
  c:person <person_uri>;
  c:charge "aggravated assault";
  c:charge "reckless endangerment".

Assuming your <resource_uri> is equivalen with your type:Conviction, your SPARQL query seems mostly correct. You can skip the ?charge2 variable though and instead just do

SELECT ?person ?charge WHERE {
  ?conviction rdf:type type:Conviction .
  ?conviction cpred:person ?person .
  ?conviction cpred:charge ?charge . 
}

because SPARQL cannot seperate the two different charge variables (afaik).

This should result in one result row per charge per person (Haven't tested it, but hopefully all is correct).

Answer by Bastian Spanneberg for What is serialisation in owl ?

2010-06-18T17:30:08Z

As OWL statements are RDF too (see OWL 2 Web Ontology Language - Mapping to RDF Graphs), each OWL ontology can be represented or formatted the same way RDF can, meaning you can, as cygri stated, serialze it into the formats stated above or even into N-Triples or Turtle.

Pros and cons for different Jena backends - SDB vs. TDB

2010-01-27T14:36:07Z

I'm still using the old RDB backend in a project and want to upgrade now to either SDB or TDB.

Which one is better when the application is expecting to be mostly writing/reading/both ? Are transactions useful (SDB with underlying DB transactions) ? Which one is better considering scale/clustering ? ...

What are your experiences with these backends and which one does fit better for which type of situations/use cases ?

Answer by Bastian Spanneberg for Pros and cons for different Jena backends - SDB vs. TDB

2010-06-18T09:55:36Z

Recently there was a thread on the jena-dev group about SDBs write performance and best practices for getting the best results. Due to some useful informations and insights there, I thought I add the link here: http://tech.groups.yahoo.com/group/jena-dev/message/44329 (that's the summary post, but the rest of the thread can be explored from there)

Answer by Bastian Spanneberg for Impractical features of the RDF stack

2010-06-17T13:40:22Z

It's more about things that bug and annoy you frequently and where you wish that the language designers had made things a bit easier for you (even if you ultimately understand why it was designed the way it is)

Blank nodes! Sure I understand they can be useful in some cases, but in most they cause problems later on. It would be nice to have a common and unique mechanism to at least have something like GUIDs asigned to blank nodes in order to make them adressable and exchangable over different triple stores.

Answer by Bastian Spanneberg for Store passwords as RDF?

2010-06-14T16:47:02Z

Currently, I like to keep these informations seperate. Most of the time, I have a database with security relevant data like the users password, and then have a URI representing a user in the triple store.

General, my use case for RDF/triple stores is to keep data/facts over which i possibly want to do reasoning or which a plan an combining with data from other data sources or in other vocabularies. I don't see this apllying to passwords.

I'm interested in other opinions, though.

Which Tools and Libraries do you use to develop Semantic Web applications ?

2009-11-02T19:10:39Z

Which tools in which languages do you use for development ? Are they open source or commercial ? Which features do you like about them ?

I think this is a good question, to gather some overview of available tools which can be helpful for beginners. I'll be incorporating the answers into the question, as they come.

Java

Jena Semantic Web Framework - a framework providing support for RDF parsing, storage and querying (SPARQL)
Joseki - a SPARQL server for Jena
Pellet - an open source OWL reasoner
Open Virtuoso - a object-relational SQL database, supporting RDF storage and SPARQL querying
Sesame - a framework for RDF parsing, storage and querying

PHP

ARC2 - a semantic web PHP framework

C/C++

Redland - a set of libraries for RDF handling and querying

Python

rdflib - a library for handling RDF

.Net

SemWeb - a library for handling RDF and querying
dotNetRDF - a library for handling RDF and querying
Intellidimension SemanticsSDK - a library for handling RDF and querying
LinqToRdf - a LINQ provider for RDF based on SemWeb

Services

Talis Platform - a service for cloudbased storing and querying of semantic data

Answer by Bastian Spanneberg for Tabling / Tableaux (in reasoners) explained?

2010-04-07T15:18:50Z

First to your questions.

How do they work in general ?

Well, basically tableaux are a set of rules for a given Description Logic, to break down complex statements in this logic into smaller and simpler pieces in order to detect clashes within this pieces. If no more simplification rules can be applied and no clash has been found in a given statement or set of statements, these statements are satisfiable.

Why are they used in reasoners ?

They are basically the way reasoners work. When I remember it right, there are reasoners trying other concepts, but most of the more recent ones rely on some sort of tableaux.

Unfortunatelly I don't know of a resources that explains tableaux calculus without relying on logical formulae. The most comprehensive source I know is the chapter on Propositional DL (PDF) of Enrico Franconis Description Logic course.

There is also a link in the material section to a paper called An Overview of Tableau Algorithms for Description Logics (ps.gz) which might be helpful.

In my bachelor thesis, where tableaux calculus was one of the base topics, I#ve worked with a paper called EXPTIME Tableaux for ALC Using Sound Global Caching, which worked for me to understand the matter, but I have to admit I needed some time to really get into it.

Finally, the Wikipedia page on the topic might also be helpful.

Hope that helped a bit.

Answer by Bastian Spanneberg for Semantic Web research Areas

2010-04-02T10:25:52Z

Although it cannot be considered as a hype topic, the semantic web gains more and more momentum, but in a more silent way. There are already questions here considering companies and success stories:

Lots of semantic sites and/or products are already in heavy use, like i.e. Twine, Freebase, DBPedia or Calais, so I don't think it can still be considered mostly academic.

Considering research, I guess it's mostly stuff like scalability and update possibilities for triple stores, mechanisms to do reasoning on large knowledge bases, versioning of semantic data and better standards. See also

Answer by Bastian Spanneberg for Open Source Sentiment Analysis

2010-02-03T15:13:35Z

I guess you didn't think about a websearch, did you ?

Simply searching for "sentiment analysis java" brings up a bunch of resources, like a related question on stackoverflow.com. From there on it's easy to find NLP/dataming toolkits like GATE, Balie or LingPipe, to just name a few (the later is partly commercial, so be careful what's allowed with the free version).

Answer by Bastian Spanneberg for How to use existing ontologies with Semantic Media Wiki?

2010-02-03T11:43:30Z

The documentation contains infos on the import of existing vocabularies. Older versions had a special page for importing ontologies, but as far as I understand it, the vocabulary import is the actual way of doing it.

Answer by Bastian Spanneberg for Ontology for university organization

2010-02-01T17:40:21Z

I don't know of a wider known ontolgy, but before you start working on your own solution, you may at least have a look into the ontology which is used to generate the data for the Lehigh University Benchmark (LUBM).

Another small project I've found out there is the University Ontology by Patrick Murray-John.

I guess non of them does fit your needs 100%, but at least you may be able to reuse some parts for your own work.

Answer by Bastian Spanneberg for Are there such things as RDF editors?

2010-01-28T15:57:50Z

You may want to have a look at Protégé, which is an ontology editor, but can also be used to enter individuals/instances. It may be a bit over the top, though, as it's a quite large and complex tool.

Answer by Bastian Spanneberg for meaning of same predicate on multiple objects

2010-01-21T12:24:24Z

I agree with Rob. Another point is, that the 2nd approach works with a blank node. This can often cause problems later on, as ids for blank nodes are only guaranteed to be unique within one triple store, so you can not work with these nodes outside the triplestore, whereas there is no such problem with the first solution.

Answer by Bastian Spanneberg for Trust in the Semantic Web

2010-01-13T11:42:06Z

Maybe "People, Profiles & Trust: On trust in web-mediated social spaces" is a nice read on the topic. It's not specially dedicated to the semantic web, but it's a result of a research project by Eric Wahlforss and Alexander Ljung - which now run Soundcloud.com - so at least it's written by guys who know much about the net.

Which are the Semantic Web success stories of 2009 ?

2009-12-18T11:05:52Z

It's that time of the year again, rankings and lists all over :)

I thought it'll be nice to see what the semanticoverflow.com community considers the "winners" of 2009 in the Semantic Web field. Which ventures/companies/products/sites do you think were the most innovative/successfull/... in 2009 ?

Answer by Bastian Spanneberg for Validate against rdf vocabulary ?

2009-12-14T10:55:33Z

I'm not completely sure, but as far as I remeber you can do stuff like that with Pellet, provided your vocabulary is written in OWL. At least you should be able to see if there are any contradictions between the vocabulary itself and the instance file.

Answer by Bastian Spanneberg for What are specific triplestore strengths and weaknesses?

2009-12-11T15:22:16Z

If you look for a comparison of query performance of different open source triple stores, take a look at the Berlin SPARQL Benchmark

Currently, the benchmark offers results for the following triple stores/RDF backends:

Jena (SDB and TDB)
Sesame (Native Store)
Open Link Virtuoso (RDF Views and Native)
D2R Server
BigOWLIM

Answer by Bastian Spanneberg for What are specific triplestore strengths and weaknesses?

2009-12-14T10:40:28Z

If you use Java, you can use Jena, which offers two storage backends.

First, there is SDB, which is based on a backing RDBMS like MySQL, Postgres, etc. The second alternative is TDB, which is a native, disk-based storage system (although not 1.x yet).

I've worked with RDB, the predecessor of SDB, and I've tested SDB, too. With the provided documentation, they are easy to use, and if there are any questions, there is a quite active community at the jena-dev group.

If you don't want to access the stores programmatical, the can also be used as backing storage for the Joseki SPARQL server, to offer a SPARQL endpoint.

Answer by Bastian Spanneberg for Who owns the content on this site?

2009-12-02T14:28:02Z

That's a goog question, indeed. I also hope the site has/will have a good policy considering the ownership of the data. A related question is "Who runs this site". Hopefully Andrew will answer this question soon.

Answer by Bastian Spanneberg for What kind of queries to use for comparing performance of reasoners?

2009-11-30T15:56:19Z

I'm not aware of a generic benchmark, but I found a paper Benchmarking OWL reasoners, published 2008.

Another place where you might look is the Lehigh University Benchmark. Some of the queries (11, 12, 13) also test some basic reasoning capabilities.

Answer by Bastian Spanneberg for How to implement GRDDL

2009-11-24T23:21:11Z

There is a GRDDL reader for Jena. Maybe u wanna use this one, or at least have a look into the sources to get hints for you own implementation.

Answer by Bastian Spanneberg for What Open Source framework can autogenerate HTML for SPARQL output?

2009-11-19T14:12:17Z

As far as I remember Joseki has the ability to process the returned SPARQL Resultset data with a XSLT template. When you issue a query in the web interface of Joseki, that's the way the resulting HTML page is generated. Should not be a too big problem to provide another XSLT script if you want to change something there.

Answer by Bastian Spanneberg for How do I build a triple store?

2009-11-04T19:06:19Z

A search on Google brought up the paper "Design and implementation of an RDF Triple Store".

If you're more like the source code reader type of guy, you mabe want to take a look at the sources of TDB, which is the native storage engine in Jena, or also at the sources of Sesame.

Update: Another place where you maybe could learn something about the topic is the BigData project. They write a lot about technical details in their blog, and it's open source, too, so you can take a look at the sources, too.

Comment by Bastian Spanneberg

2010-07-08T14:07:26Z

This seems to be a problem in other triple stores, too. I encountered the same behaviour with a DB-backed Jena triple store. i'm curios if anyone already solved this ...

Comment by Bastian Spanneberg

2010-06-25T09:43:13Z

Jena is a great toolkit for sure, but has no support/tools for RIF.

Comment by Bastian Spanneberg

2010-06-17T10:08:30Z

just for completeness, the link is topicmarks.com

Comment by Bastian Spanneberg

2010-06-17T10:05:21Z

thanks for the "see reasoning hint", that's indeed a nice look inside the mechanism of the website.

Comment by Bastian Spanneberg

2010-03-11T21:23:00Z

i totally agree. we've had several similar cases where blank nodes also caused these types of problems. we also use UUID to generate URIs where possible. the downside is that you need some code in your app to take care of this, but imo this pays off later in matters of data portability.

Comment by Bastian Spanneberg

2010-03-03T09:18:04Z

@dpinn: ORDER BY on xsd:dateTime works just fine.

Comment by Bastian Spanneberg

2010-02-22T15:10:37Z

Nice and comprehensive tutorial. Thx for the link.

Comment by Bastian Spanneberg

2010-01-26T14:40:30Z

but these are no Q&A sites, are they ?

Comment by Bastian Spanneberg

2010-01-22T11:04:26Z

sure, but still you need a mechanism/algorithm which produces the URI for this node. so you may have to change your business logic to fit this, although this part does not really belong there (i had a similar problem with blank nodes in a project, which gave us a lot of headaches).

Comment by Bastian Spanneberg

2009-12-01T19:56:34Z

Glad I could help you. As far as I understand, only the queries 11-13 test some reasoning capabilities, but it's a start.

Comment by Bastian Spanneberg

2009-11-17T17:05:20Z

I had a look at the pdf, but still I cannot figure out which namespaces you mean. Can you clearify your question ?

Comment by Bastian Spanneberg

2009-11-13T13:44:05Z

I have Segarans book too, and it rellay faces the topic straight forward. Can definitly recommend it, even if you're no Python programmer (the examples are in Python mostly), which I am not for example.

Comment by Bastian Spanneberg

2009-11-09T14:51:14Z

i also think that a low-cost solution is impossible in most cases. everytime i've worked for a customer were i could see a great benefit for him using semantic web technologies, it was not only a question of data layer, but also a lot of program logic and the likes.

Comment by Bastian Spanneberg

2009-11-07T16:04:53Z

the list of citations is indeed another good starting point to look for other stuff on the topic of reasoning. found a bunch of interesting papers there.

Comment by Bastian Spanneberg

2009-11-07T14:18:49Z

give the site some time. it just started, and i think there are already nice discussions going on. sure it'll take some time to build a userbase, but i still think it's worth to have such a platform for the semantic web topic.