Author: leobard

https://www.leobard.net/blog/about-leo-sauermann/
Posted on

Showing gnowsis.com at CeBIT – see you there?

So, after many months of setting up a company, we have our second public appearance: gnowsis.com is going to show up at CeBIT. We will be giving demos of our first product “cluug.com” and give away alpha accounts for selected users. There will also be a great video about it.

We have been working many months to get a minimum viable product running. That means, we are going to offer first the core feature of semantic personal information management: links. From the vast amount of things we learned from the NEPOMUK EU project, the first thing that is commercially available from us are links between websites. That gives you the chance to try out the thing online and us the chance to give you more features every month.

So, to learn more: come to CeBIT 2010. I and Martina Gallova are going to be at the gnowsis.com booth as part of the DFKI booth in Hall 9, booth B45. We will also be giving a future talk on Thursday, 4th March, at 17:00 in Hall 9. And we hope to be able to show people what we do at Bitkom’s innovator pitch (you can also wish us good blessing for winning a price there).

I am very happy with the forthcoming of the company so far. Everything took a bit longer as expected, but the results are nice. Between December and January, we rehauled the user interace (kudos to Bernhard Schandl and the development team for leading this). We are negotiating partner deals with interesting companies to bring the semantic desktop and cluug.com to you as soon as possible.

So, I won’t go into details, this leaves plenty of things to talk about at our booth at CeBIT. If you want to meet us – please please write me a mail before. We are going to be very busy and it would be good if we can make an appointment beforehand. I can also organize you a free entry ticket for the meeting, if you need one.

Especially if you are a journalist of professional blogger, it is always better to plan meetings beforehand. At the DFKI booth there is always press around and politicians creating buzz, if you want a calm moment, maybe take care and book one.

Of course, there will be Mozart Choccolate Balls at the booth for you (my traditional talk give-away), we also give away some buttons, but most important – very very rare Alpha User invitation codes.

(btw: ESTC 2009 and getting a price at innocation camp there was our first appearance)

Posted on

Accepted at LIFT-Austria Conference

I submitted a short proposal to speak at LIFT-Austria and got my acceptence mail!

See you there on March 18–20 in Vienna, its going to be interesting. My talk will be about something around “tools for thinking – how semantic personal information management is going to change our way of thinking”

Looking forward to the event and to meet others in the field, and to drink a beer. If you are also going, ping me.

Posted on

Input for a possible RDF 2.0

There are activities towards updating the RDF standard.

Here are my thoughts on what problems and solutions we have:

Leobard’s thoughts about needed changes to RDF.

Problems:

1) Reification is not an aesthetically appealing model because it forces the triple/statement structure on quads. Therefore it is not used much and discouraged by some “named graph” enthousiasts. Nevertheless, the need to identify and annotate single triples and their values is there.

2) rdf:value, datatype, language, and reification all address the same need and are redundant.

3) The relation between a web resource (i.e. a web page in html) and the RDF document (named graph) containing the RDF data of the web page can NOT be expressed with the RDF standard. There exist various, scarcely documented methods such as the HTML Header tag “” or 303 redirects, or content negotiation. Some of these methods are described in “Cool Uris for the Semantic Web”). This has been causing personal bellyaches for me since editing “Cool Uris for the Semantic Web”. It is not aesthetic as this central feature of linked data and RDF can’t be represented in RDF.

4) Statements about reified triples must be possible for sets of triples.

Suggestions for Solutions: [syntax: problem->solution]

1)->S1) On the core level of RDF, add an URI identifier to a triple. Let Serializations allow to add this URI to the triple. Add a triple identifier to the core of the spec and APIs.

2)->S2) Deprecate rdf:value.

3)->S3) In RDFS we already hint at HTTP dereferenciation and linked data in rdfs:seeAlso and its subproperty rdfs:isDefinedBy. In foaf we have foaf:homepage that links a resource to its web page. In SKOS we had skos:isSubjectOf (but it was removed) I propose “”” rdfs:describes a rdfs:Property; rdfs:domain rdfs:Resource; rdfs:range rdfs:Resource; rdfs:comment “The subject RDF resource is metadata for the object document.” “”” . This solution seems to add problems though, as the relation between document and resource is dynamic and ever changing.

3)->S3.1) Leave it as is. The problem of linking between HTML and RDF representations is on the level of HTML and not on RDF.

Posted on

Mannheim Lawyers go beyond law – sue twitter user “mannheim”

The german city “Mannheim” is suing the twitter user “Mannheim”, a guy living and working in Mannheim. CAN YOU HAZ TWITTER?

read the original post:

Mark has now received a letter via registered mail from the City of Mannheim, which says that he must sign the letter and give up the Twitter account, or suffer the full force of a city’s legal team as they try to drag our a**es through the courts.

In my opinion, the city of Mannheim can fuck off. If they forget to register their own name on twitter, they are clearly years behind. And they will always be too late in all the good services, so if they start doing this with Twitter, where will it end? The german legislation allows you to sue for DOMAIN names, which is kind-of-ok, but with web 2.0 accounts? Anyway, I am with the small David here against Goliath.

Posted on

A late NEPOMUK deliverable: the personas

In the NEPOMUK EU project, we have created standards and implementations for the semantic desktop. Based on which assumptions?

To get an understanding what people do at various companies and what support they need, a set of personas was created. Based on interviews with real people, a persona is a fictitious person that represents a user group.

Here is Claudia, she and her colleague Dirk were the two most popular personas within our project group:
claudia stern, a persona

The personas helped us to think about what the software must do for the users, demo the software, create prototypes, and create test data for unit testing. As they helped us, maybe they can also help you, so I asked around in the consortium if we could publish them, and we could, so here they are:

A definition was given by the authors from the Human Computer Interaction Group at KTH:
A persona is a fictitious person that represents a user group. They are based on users studies on real people.

Personas are a detailed description and a visualisation of the users. They have a life, goals and scenarios where they fulfill their goal. They help us to focus on the users during the design and give all stakeholders in the project a clear picture of the users’ needs and requirements. Everyone in the project has the same view of the users and personas are also a constant reminder of the users.

When personas are used in the design work and they make it easier to design for them. They “depersonalise” discussions on functionality and allow the designers to focus on designing for the personas.

you can also find the personas in the list of deliverables:
nepomuk.semanticdesktop.org/xwiki/bin/view/Main1/Deliverables

p.s. The personas were so good, we continued using them at the NEPOMUK KDE Workshop. There, the story continues with news on Claudia’s private life. In fact, she is having a wedding with her long-term friend Berit in Holland! Read the fascinating news and N3 files yourself…

Posted on

Making the Tevion IWR 394 work with linux server

Ingrid gave me the great Tevion IWR 394 as a christmas present.
My first radio with rj45 jack! Bliss! (it also does wifi)

We named it “Adam“, our iPod is called Eve, so they are a good pair.
Adam has UKW FM radio, playback of mp3s on USB drives, internet radio (the first station I listened to is, of course, slay radio), management of favorite radio stations via the wifiradio-frontier.com webservice, Wifi connectivity, LAN RJ45 connectivity, playback of your own media collection using microsoft windows media player 11 media sharing, and he also has an alarm clock.

See it in action:
"Adam" - Tevion IWR 394 - slay radio!

now the problem is of course: how can Adam playback my iTunes library from my Apple iBook G4 “Eden”? I have a backup of this iTunes library on my linux home server “Franse” – so hosting some magic server on Linux should do it, right? right.

My device is a Tevion IWR 394. I can’t find anything about it in the interwebs (crazy?! nobody bought this before?), so it is probably a branding of an asian white-label product. The menu shows a software “ir-mmi.arts.ven6-jupiter6.1_V1.5.6.21635-1A19”. It connects like a baby to a breast when using UPnP and my Windows XP laptop. FM/UKW radio, usb, also work easily.

Background: accoring to microsoft help of windows media player 11, the following ports should be opened to have media streaming working:
1900 UDP, 2869 TCP, 10243 TCP, 10280-10284 UDP. There is a wikipedia page on the Media_Transfer_Protocol which is to connect your media library to your USB device. For network media streaming, UPnP_AV_MediaServers seem to be the right place. There is a upnp server comparison table on wikipedia also.

My own requirements:

  • host a lot of mp3s from my iTunes collection (ideally, also host my DRM protected itunes files, but thats probably impossible)
  • be piss-easy to install on a headless ubuntu
  • import my iTunes playlists and song ratings

How they meet this:

ok, having found mediatomb as my “easy way to go”, I try
sudo apt-get install mediatomb-daemon
fine! I see it installs it as daemon on rc2.d also, so it will start on system boot.
The various install instructions (ubuntu, official) go on:
sudo nano /etc/mediatomb/config.xml
hm, seems I don’t have to change anything in config.xml, that looks fine as it is. The password and username, ok.
I go to my server using
http://franse:49152/
(franse is my server name. it won’t work at your place)
Ok, I see a very minimal user interface, fine. Somehow I find the folder I want to share and click “+”, it starts indexing. That makes the mediatomb process jump to 80% cpu time on “top” for some 10:50 minutes. good.

But – most amazing – the files show up on my Tevion IWR 394. Life is good!

It does not read/interpret my itunes playlists, though. Googling for itunes playlists in mediatomb does show an empty graveyard. Ok, but its possible to convert an “iTunes Music Library.xml” to several .m3u files using itunesexport, a java tool (here java bytes me again). Sadly, the ericaugherty sf version causes an out-of-memory even with -Xmx300m, so thats a bit annoying. There are zillions of shabby scripts out there to do it, but most don’t work, maybe this one does, as the author says. It does! But it needs some tweaking for folder names, which is done using the “-d” parameter. So the command is something like this (note the trailing shlash at -d) which I put into updateplaylists.sh
#!/bin/bash
python /home/leobard/local/bin/itunes2m3u.py -d “/home/media/music/iTunes/iTunes Music/” /home/media/music/iTunes/iTunes\ Music\ Library.xml

(at this moment, my newly discovered and totally weird emap.fm internet radio station plays the amazingly happy disturbing song “It’s so chic to be pregnant on christmas“, greetings to the pregnant-women-we-know B+S)

It seems importing m3u playlists into mediatomb is not out of the box. It goes something like this:

  • check the playlists are allright. do they point to the right path? are the characeters fine? do the directories exist?
  • in my case, the playlists are in /home/media/music/playlists folder and contain lines that look something like this: /home/media/music/iTunes/iTunes Music/Christoph & Lollo/Schispringerlieder/04 Mika, Du Saufkopf, Hast Du Wieder Verloren_.mp3 – that means, absolute paths do work fine for me
  • I import the playlist using the explicit “+” button for the filesystem folder in MediaTomb. Other things didn’t seem to work, especially the autoscan did not work.
  • I thought about script the adding of the playlists by using the mediatomb –add command … but that is not optimal. It starts a server and adds the file, but the server is not shut down. Not ideal for a command line thing.

So to sum up: it would be good to have a DAAPD and upnp server in one, which could read iTunes playlists. BUT I am equally happy with my Tevion IWR 394 “Adam” and mediatomb.

Posted on

IT-City Vienna

“Vienna will be evolving into a city of knowledge, away from repetitive work to a diversification of simple manual labor and high-skilled jobs. This already happens and must be aknowledged by economic politics.”

somehow like this go the introductionary words of an Article titled “IT Stadt Wien” in “Information Professional” 4/2009, the magazine of the UBIT IT guild in Vienna.

What caught my eye is the wish for more software products coming from the Vienna region. As we have a lot of semantic web companies around, that could potentially also mean: how can we do semantic web products in Vienna? Currently, a lot of software is bought from international corporations (lets call some names: SAP, Microsoft, Siebel, …) or done from scratch in software development projects. What we need are more product-producing software companies in Vienna. Especially we need them because by going for a beer with people working in productive companies, others can learn how to run a business (the author of the article put it “the needed knowledge is not here…”).

My 2 cents:

  1. yep, when I finished university, the choice was “install microsoft windows server 2000 for the rest of your life”. or try something on yourself – which I did because the alternative is boring
  2. what about Altova?

To my knowldge, Altova is the only company in the Vienna region that sold a PRODUCT. With PRODUCT I want to say: they have a software that has the quality, marketing, features needed to sell it worldwide as an off-the-shelf product. They opened up offices in Beverly, MA, USA also, as far as I know to improve sales in the US market (did they also move development and taxing?). Who else did this in Vienna?

Is Altova open to share their knoweldge? I would be the first one lining up to learn from them, being in the situation to try to do a similar feat.

Another point: the wko magazine didn’t mention article author’s names – for a magazine this is a no-go, I want to know who wrote what and to whom I can address letters. Add names!