Recreational computing

There is so much going on recently in DERI/eLITE/Corrib - so many things I would like to write about ... but I can't. At least until all those strange IP policies will get firm borders stating what we researchers can tell and what we should not tell.

Until then, trying to avoid any NUIG IP policies land mines I might stuble upon, I can only tell that I came back to unlimited-fun-generating activities: research & developement. I have managed attract a group of skillfull researchers to take care about each of the projects I set up some time ago: JeromeDL (Tomasz), FOAFRealm and HyperCuP (Sławek), MarcOnt and a very new one S3B (Adam).

Now I can relax slightly from some the management responsibilities and spend some time on "recreational computing". For me it is a combination of all the I really like: maths, user interface design, web programming/prototyping, and ... inventing.

I hope I will be allowed to publish some of my recent ideas and prototypes. Untill that time, unless you are from NUIG/DERI, sorry ... you got to trust me - I am having great fun (although it employs working 10-12h/day)

Beware of CVS tags

It took me a while before I realized what went wrong.
Recently we have finally announced a stable version of JeromeDL 2.0. I was so happy about that.
So the only thing that was left was to tag our current sources with stable tag and keep on rolling with further development.
All went well, until Mariusz was trying to update library.deri.org server (and other) when we realized that something was not right. It occurred that some time in the past we also used stable tag to annotate stable version at that time. Since then a lot of files were refactored, moved, etc. But they still retained the stable tag - which caused why they got into our recently updated servers when cvs-updating to stable tag.
The only way I found to fix that ASAP was to go over all files with errors, warnings, and those we knew were removed and ... remove the stable tag.
Now everything works fine. But, the lesson was learned - be careful when re-using tags in CVS....

Geotagging made easy

I was always jealous for all those people that had their photos annotated with GPS positions - especially since I travel quite often and after some time it would be very nice to see where I've been precisely.

Now I can ! With Flickr geotagging - that is as easy as you can imagine.

I do not have to take care about GPS, geo-positions, etc. Just drag-and-drop - and you can easily place your photo anywhere on the Planet :)
Plus, other people can not only see where it was taken (click the map link in the metadata section of e.g. this photo on Flickr page) - but can also know the name of the place where it was taken.

Cool, isn't it? Now you can see where is my in-law's garden where this photo was taken :)

And, btw. I managed to geotag about 1000+ photos so far (more coming soon) which is just a fraction of what has been geotagged by the whole Flickr community within last 24h.

Calling all del.icio.us people

I am just in the middle of research on how to better represent tags than just with tag-cloud. My prototype is up-and-running, however, I would like to introduce some more smarted clustering and filtering techniques.

(help me to speed-up my research)

In order to do so - I need your help - all del.icio.us people - if you want to experience better way to browse your del.icio.us posts - send me your del.icio.us IDs (no passwords :) - so that I can add you as friends.
It will allow me to gather more data so that I can improve my current algorithms.

I appreciate any help - and I promise to give you a credit on the project web page and source code once I will announce this component.

Thanks in advance

Not so wet in Galway, afterall

(Galway Eyre Square by Kopretinka)

Jacek says it still rains in Galway - but I bought a bike 2 weeks ago - and so far I did not managed to get wet while biking to/from DERI :)

technorati tags:galway, wet, biking

Blogged with Flock

KDE and Beagle

I have been wating quite a long time to be able to use Google Desktop Search-like in my Linux box. There were two candidates: Beagle and Kat. Until Fedora Core 5, Beagle was crashing all the time, and was hardly performing any indexing. FC5 added Beagle as a part of distribution.
The last thing I needed was some KDE front end - and Kerry is nice enough.
But there was one flaw - I do not like neither Thunderbird nor Evolution. And the way Kerry tried to open emails was resulting in new emails with the content I wanted to read as attachments.
The solution is quite simple - write own wrapper for KMail to handle that. Since I did not know at first wheather the content of the email is piped to the program or just referenced, I make a small script that handles both ways. I hope others will find it usefull.

#!/bin/sh

if [ ! $1 ];
then
      TMPFILE=`mktemp -q /tmp/emailpreview.XXXXXX`
      if [ $? -ne 0 ]; then
           echo "$0: Can’t create temp file, exiting..."
           exit 1
      fi

      while read -u 0 TMP
      do
              echo "$TMP" >> $TMPFILE
      done
else
      TMPFILE=$1
fi

/usr/bin/kmail --view "file://$TMPFILE"

if [ ! $1 ]; then
      rm -f $TMPFILE
fi

DMoz vs RDF Repository (Sesame)

What can you do - there are just some days that just should not happen. I think one of them was when I decided to integrate DMoz ontology into JeromeDL and FOAFRealm. It all looked so harmless - especially when I got (small !) part of the ontology from Andreas. I build the whole mind model around that, finally even set up a JOnto project to deliver unified API to handle taxonomies. And ...
I decided to download DMoz RDF, or what ever they claim to be an RDF :( It took me some time to realize what was wrong. And eventually I got some help from Hee Chul and Krystian with nice converting scripts. I though that it was the end of the problems - I had a sample RDF (a true one) that worked. And a real RDF version of full DMoz ontology. But it was not the end of the problems :(
I decided to upload the 800MB RDF-DMoz file to Sesame. But after a couple of hours of waiting, 100% CPU usage, almost 80C CPU temperature of my laptop, I gave up.

Daniel suggested I should just point to the RDF file and make the memory repository. Well - it went quick that way - "out of heap memory" error :( Later I took the "divide and conquer" approach. Cut this 800MB file into 10 smaller. First one got uploaded very quickly (relatively). And so, encouraged by that example I started uploading the rest 9. Each next of them was taking much much longer to be uploaded, until the 10th one that obviously must had make Sesame hanging - as there was no progress for the whole night (I went to sleep btw).

I cleaned the repository and uploaded only the first chunk again. But trying to use it - with browsing or SeRQL querying was way to sluggish. Finally I came to my brains and "slimed" the DMoz RDF removing (with modified Krystian's script) all information that was not defining dmoz:Topic or using dc:title and dmoz:narrow{12}. Luckily I got 200MB RDF file that went smoothly into Sesame.
And now JOnto-DMoz is finally kicking the ass :)

[I will upload the scripts and the final RDF file to jonto.sf.net soon]

Carnival 2006

pict8974
Originally uploaded by skruk.

And so we made it. Everyone said it would be a shame to be in Rio during the Carnival and miss the Samba Schools Parade.
Luckily I have even managed to convince myself to take my precious camera with me, shot almost 600 photos, and bring them all home safe.
The funny thing is that what ever people say about Brazil, although you can get easily paranoid with respect to ones safety, Carioca are really nice people. During the parade the only hostile people where two geyish Englishmen trying to quarrel with everyone (including myself) that was trying to get too close to the fence (as they claimed to be the only ones that should occupy that place and shot photos). Weird, isn't it?
Any way - presentation was magnificent - we were only sorry we could not make it to the end - after 5 school we decided to go home (it was 6am BTW) - as I could not see a thing (600 photos / 6h can make your eyes hurt - believe me) and Ewelina was tired as well.

DigiMe - personalization vs privacy

Contemporary recommender systems derives most of their power from the profiling information on the user. The problem that usually arise is that by amassing infomation about users - recommender systems became a threat to our privacy. The more intuitive and effective search/recommendation system we want to get - the more information about ourselves, our fancies, our friends and foes we need to provide to the system.
Since I always believed that this is to much to pay for just a recommendation - I have came up with the Social Semantic Collaborative Filtering idea that users themselves have control over their profiles. The current implementation in FOAFRealm/D-FOAF project is still far away from what I would like to achieve.
Our aim is to deliver a true identity management (DigiMe) that will not only identify a user in the Internet but also will provide better trade off between personalization and privacy.
I have just came across the presentation DoDEA_Emerging_pt1.ppt
that presents similar opinion to mine in that aspect: we must build identity management system that will allow personalization of services keeping the privacy the same time.

Music for free (almost)

Recently I have convinced myself to buy a MP3 player with 1GB and radio, just to keep me entertained on the plain when I cannot use my mobile phone and in any other occasion where it is going to be more handy.
But soon I have come to the point where all my MP3s generated out of my legal CDs were no longer cool and entertaining anymore. Buying a new CD is not an option this days, so I went Internet-shopping.
Luckily "legal mp3" query in Google led me to http://www.tunster.ru/ which turned out to be the coolest (and cheapest - please correct me if I am wrong) places.
An average mp3 costs around $0.10 which means that I can stuff my MP3-player for around $15. And the quality is really good - 192 bit rate.

Polish community in Galway

Today was the first polish mass I attended in Galway. Great thing that we can finally have one. But I was also said that not so many people were there - probably because that did not even know about it.
So to change this situation I decided to set up a blog [http://poloniawgalway.blogspot.com/] where polish community in Galway notify others about events like that.

How to advertise a school of motoring

How to advertise a school of motoring
Originally uploaded by skruk.

Ever wondered how to do that? Try to squeeze your car in a small alley, park it inch beside the fence ... and there you have it - it just screams: "no other school will teach you how to park that good"

DDict - Desktop Proxy for Online Dictionaries

For a long time I was using a nice desktop interface of the dict.org service for KDE. But it was of no use when trying to come up with a translation from Polish. While there is a lot of good online dictionaries - none of them had an similar proxy for the linux desktop.
I have decided to take appropriate action - one night of hacking and I have developed DDict - Desktop Proxy for Online Dictionaries [http://sf.net/projects/ddict/]. This small tool written in Java/JDIC resides in System Tray allowing user quick access to a number of predefined (and easily extendable) list of online dictionaries.

Check out the ready to use packages at http://sourceforge.net/project/showfiles.php?group_id=157520 and screenshots at http://sourceforge.net/project/screenshots.php?group_id=157520

History of MarcOnt

I have just realized that although so much has been said about MarcOnt in many places, I have never said the full story of that initiative. It all stared when I came for the interview to Galway in April,2004. I meet a lot of interesting people from DERI - current colleagues. Among them I meet Anna. We have briefly discussed what we are working on. She told me that since the ontology has to be build based on the community agreement - she is building a portal where a community/e-society can collaborate on the ontology. At that time I had problems with librarians accepting the idea of the Semantic Web. The only remedy was to create an ontology based on their favorite MARC21. But how to do that? Then it struck me - we can get librarians to create the ontology for themselves. But we need a smart process, with negotiations, collaboration and versioning. And this is where MarcOnt Initiative and MarcOnt Portal were born. It was somewhere over Germany on my way home from Galway (through Frankfurt) ...

Rainbow

Rainbow
Originally uploaded by skruk.

No wonder Ireland is such a rich country: Rain -> Rainbows -> Irish Pixies + Pots of Gold -> Lots of Gold !

A hole

A hole
Originally uploaded by skruk.

I am just curious - is this the way to cover the hole or to warn pedestrians about the danger?

Do not fly to/from Bydgoszcz

When 3 weeks ago I was land in Gdansk instead of my destination - Bydgoszcz - it was kind of funny, since I got a coach that dropped me out in my hometown on the way from Gdansk to Bydgoszcz. The weather was horrible - first snow - and we appreciated the care Ryanair took to handle the problem of its passengers.
But what happend today was a completely weird story - the weather was good - not perfect - but you could see clearly for miles - so we were sure there were going to be no problems. How wrong were we ... the plane decided not to land in Bydgoszcz due to fog (what FOG?!) - and fly to Poznan instead - leaving all of us on the ground - helpless :(
What came next was like a horror - of course we were offered to get our money back but when you book your tickets early - the price then and for the nearest flight was like 1:10 - nice solution. Or to re-schedule our flights - problem was - 200 people waiting in an endless queue - hearing that the next available flight was no sooner that in a weeks time.
In addition if you were to have a flights connection to e.g. Shannon or Derry like some of us did - you were in even more trouble - no help at all - no money return for the second flight - unless your private insurance covers that (mine claims it does - I hope so).
To cut down the number of passengers waiting to be servered - we were given small pieces of paper with Ryanair web site (wonder how I book my tickets without it) and a phone number to Dublin where we were supposed to be served.
So I decided to check the flights myself immediately - there was Orange.HotSpot and since I am used to using airport WIFI services - I was trying to connect to that one. To my surprise Orange do not provide online service to buy the access time - the only way to do that is to buy one of those scratch cards- but ... surprise ... you will not find one on THAT airport :(
Finally I got home - and find two possible connections I could you next week - reasonable expensive - but as you could suspect - not provided by Ryanair. I decided to do the last thing I could - call the phone number we were given and arrange reimbursement - to my surprise the line was dead with information stating that I should call between 9-14 and 15-19 Mon-Fri - since it is already Friday, 9pm - there is nothing I can do about that, Damn!!
Take my advice - DO NOT FLY TO/FROM Bydgoszcz