personal weblog of Leo Sauermann
I wondered “how does iTunes store cover images?” but was too lazy to use a conventional keyword search engine, so i searched for a natural language processing search engine and found brainboost.
it works. the returned documents did answer my question. I love the web.
previously I blogged about a hack by crschmidt. I have to add that crschmidt continued work of Masahide Kanzaki.
crschmidt did something that sounds interesting: glue code to bring together masahide ideas, XSLT and flickr. sounds very interesting.
have to look at it! 🙂
someone said to me in amsterdam that I have good tourist karma. Today I have good web karma.
from boingboing.net I wandered to web zen which inspired me to go to leaveyourmessagehere.
the idea at this place is simple: a text-input box and a button “leave your message here”.
under this creativity pressure I wrote this:
come to kaiserslautern. I am there and I miss you.
after pressing the hardest button, I saw the last message there (5 hours ago)
Thinking of you P……..do you miss me?
web karma: your spirit is so in tune with the web spirit that you react to things you conciously cannot know.
michael notes that one could use http redirect and that clients should update their links when finding one of these.
After a permanent http redirect of a uri (hence also a semantic web uri) all triples and so should be updated.
some clients forget to do this and show the new data but keep the old uri.
practical rdf failure:
I moved my foaf file from gnowsis.com/leo/foaf.xml to http://www.leobard.net/foaf.xml – that has been a year ago.
still many foaf files have their seeAlso set to the old url. So I wrote to eaon/Michael Zeltner to update his foaf file at
http://zine.niij.org/data/foaf
he replied that his foaf file is now:
http://zine.niij.org/data/about-me
so we both moved our files. and crawlers / search engines like this here for eaon or this for me or that for me still have the outdated data.
a nice approach for identifying thigs with uris (known as the uri crisis) is the taguri.
note that taguris are not wise to be used with things that have already a http:… uri. And in my opinion, using any existing wiki system to make a URI has the same outcome as taguri and is easier to handle for the lesser deamons. (we, the high arcane sorcerers of semweb, have a domain, know how to use a date and can live with a uri that points to nirwana on the http space. lesser demons enter uris into browsers)
but despite these flawbacks, taguri is a nice idea.
thanks to andy seaborne for mentioning it to me.
an interesting thing happening in Austria is the Semantic Web School. The aim of this school is to preach the gospel of the semantic web to the masses, especially industry, government and end users.
The website has information about the school (in german) and the offered curriculum. They offer a set of one day workshops about several topics. The events are spread over several months (see navbar on left) and can be booked seperately. Cost is ok, compared to the quality of the speakers it is surely worth it (and regarding the good connections you get to the other industry & govn attendees 🙂 At the moment, courses are ongoing.
A very nice thing about the website is the good collection of information about the semantic web at large, they collect much info. They also asked their speakers to contribute there, so expect a quite well kept library of semantic web stuff.
Look at the right side of the nav bars to find the good links.
and yes, I am also giving a workshop there, about the dirty and good RDF tools: redland, jena, sesame, protege, gnowsis. I am looking forward to this, its on 25th may, you can still book to watch me.
Again gnowsis has advanced and the 0.8.2 release contains some interesting features, including a Joseki and SPARQL server and some bugfixes.
We are aware of the current movement towards sparql and have added SPARQL as a protocol to gnowsis using the ARQ/Joseki implementation. This can now be used in parallel to the existing URIQA interface, that has already been used by developers to talk to gnowsis.
Basically, gnowsis is a Semantic Desktop framework that integrates heterogenous data sources (files, outlook, imap servers, etc..) into a queryable and linkable RDF framework. It is programmed in Java and extends Jena. You can use it as a very big code sample of how to use Jena, install it and browse your own filesystem as an RDF graph, extract Concise Bounded Descriptions of resources and be happy with all the RDF.
Gnowsis is released under a BSD license and you are free to extend it.
The 0.8.2 release is mainly a bugfix and enhancement release of 0.8.1. We will now distribute gnowsis more frequently and move from our private cvs to a public visible subversion soon.
download gnowsis here:
http://www.gnowsis.org/Download
Changelog of 0.8.2:
* a new Joseki Server. Joseki is added using the hot joseki3 version from the latest CVS from sourceforge. Expect to make SERQL queries. documentation follows
* the service concepts has been improved
* gnowsis uses now Jena Modelmakers in central places. This allows to configure the search and central repository services to run either on database or on file models, in the future also NG4J
* Client interfaces made better.
What misses is decent documentation of the new features. If you are willing to contribute, please contact leo@gnowsis.com
If you are interested in gnowsis, don’t hesitate to contact me,
may rdf:Resource be with you.
Leo Sauermann
note: jena schemagen omits entries for RDFS:classes when you do not pass the argument –rdfs. So when you search for missing class entries, try it.
version jena 2.2
tricky tricky, it used to do it, I think.