Friday, April 28, 2006

Any medd for your girl to be happy!

Last week we learned that 60 billion emails are send worldwide on each day, 97% of which are SPAM (assuming that the numbers from Germany's biggest internet provider are an adequate sample).
To cheer you up after this shocking number: the original Monty Python sketch that was the reason unwanted commercial messages where christened "SPAM". The three and a half minute sketch manages to mention SPAM more than 130 times.

CfP update

Cfp update: cleaning up my inbox ... some CfP's that aren't expired yet (some barely)

Monday, April 24, 2006

16000 Dead Germans / Year

According to official data 16000 Germans die every year because they where ordered to take drugs that interacted in some way not anticipated by the doctor. 100 000 are send to hospitals for the same reason. Many of these cases are caused by well known interferences of drugs, not anticipated because the doctor did not know about them or the doctor did not even know all the drugs a patient was taking.
It is believed that electronic storage of medical data and AI technology may reduce this number - and such a system that would give a chip card to every German was supposed to be in place by January of this year. The chip cards would be able to store some information (prescription of medicines and data helpful in emergencies ) and would act as a key to centrally stored information (such as XRay pictures).
Sadly creating such a large computer system in an area where data security is of high importance is very difficult and now delayed for some unknown time period. I just read the not even the specification is completed :-(

Wednesday, April 19, 2006

(German) Computer Scientists Not At All Influential

The magazine "Cicero" compiled a list of the 500 most influential Germans. The list tries to reflect who the mainstream media asks, cites and interviews. I read the entire list to find the most influential computer scientist (my bet was on Wahlster). The result surprised me: There wasn't a single computer scientist! The number of other natural scientists is very small as well - most of them medics or physists; there was not a single mathematician.
When trying to explain the world, German media apparently turns to writers, philosophers, journalist, historians ... most of them male and old (average age of the top 100 is 66 years, 91% male). Anyone still surprised that the German public opinion is sceptical when faced with new technologies?

The International Day Of The Backup

Technique and computers in particular have become a part of our daily life, but since they are quite new phenomena they are not yet many customs, rules or rituals governing their use - which is a shame, customs and rituals are great ways for society to spread useful knowledge and encourage beneficial behavior. Hence, let me propose one IT related ritual: the "International Day Of The Backup". The day where you burn all valuable pictures, your $500 iTunes collection and important data on DVD and store the DVDs at your friends houses (or bank vaults). (yes, I am serious. Sure, its not as important as "Womans Day", but it would really be a great way to remind the usual home user to do backups)

Monday, April 17, 2006

Microsoft Academic Search

Microsoft recently revealed the beta for their academic search - an attempt to compete with Google Scholar, which has been around for a year or so (no, I had a look, Yahoo is not yet offering something similar).

I use Google Scholar quite a lot, but immediately liked the interface of "Windows Live Academic Search" more. Their interface offers one critical feature I've been missing at Google: a way to directly access the bibtex data of the papers I'm looking at; with Google Scholar I often needed to do another web search to find the data needed to correctly cite a paper. On the other hand - at least in my area of Computer Science - Google Scholar seems to index a lot more papers, which in the end of the day is still the most important feature for a search engine.

Saturday, April 15, 2006

Free Our Data!

From the Guardian:

Our taxes fund the collection of public data - yet we pay again to access it.
Imagine you had bought this newspaper for a friend. Imagine you asked them to tell you what's in the TV listings - and they demanded cash before they would tell you. Outrageous? Certainly. Yet that is what a number of government agencies are doing with the data that we, as taxpayers, pay to have collected on our behalf. You have to pay to get a useful version of that data. Think of Ordnance Survey's (OS) mapping data: useful to any business that wanted to provide a service in the UK, yet out of reach of startup companies without deep pockets.
This situation prevails across a number of government agencies. Its effects are all bad. It stifles innovation, enterprise and the creativity that should be the lifeblood of new business.
Read the entire piece, after that you may want to sign here.

Friday, April 14, 2006

European Grand Challenge

Just read about an "European Grand Challenge" called the 1st European Land-Robot Trial. The usual European stuff: rules more complicated, not one winner but many, not an competition but an evaluation of existing technology, no cash reward ... the track is much shorter but more difficult, the race allows everything from fully autonomous to remote controlled vehicles. The only thing not "typical european" is, that it is still a military competition. I had hoped for a European Grand Challenge spin off in the civilian sector, focussing on the problems of "social navigation" - with many cars, children, people on bicycles, traffic lights ...

Thursday, April 13, 2006

Spring!

Yea, could have taken the usual spring flowers, but though a couple of weeping willow's is nice as well.

Query Response Time

I often hear people say that inference based question answering systems will not be able to compete with current internet search engines because people are now used to less than a second query response time. That people will not be willing to wait one minute - even if the answer is better.

In a way I agree: With respect to traditional web search this may be a realistic hindrance to the introduction of more knowledge intense algorithms. But on the other hand, we can see that the internet will be accessed by devices other than the traditional desktop computer. Devices such as mobile phones, for which expectations on search have not yet formed. Just consider this prototype from Google

Searchers call a Google supplied number, where a prompt asks them to "Say Your Search Words." After a minute, the results are returned and, in the demo, the searcher clicks a link that leads to them. In a real-world application, the results would presumably be sent to a cell phone screen or in-car system.

Users do not yet have expectations on how long it will take to send them an SMS with a result after they spoken a query into their phone. They will not even realize that their car navigation system waited two minutes for the answer to the query for a personalized collection of podcasts with commentary about the area they are travelling through. They will still sleep while their eBook waits for the personalized newspaper (the eBook asked the alarm clock when it has to start).

Tuesday, April 11, 2006

A First Look At Quaero?

You remember Quaero, the (partly) public funded German-French cooperation to create an internet search giant? Two large german IT journals (ct and IX) are reporting that you can already see some version of their software at www.exalead.com (made by the French part of the project, the German part hasn't started yet). The site looks nice, has an innovative interface and some cool features (I liked the simple way to restrict results by location and/or language). But is this a version of the Quaro Software? I don't think so - Exalead is a project partner of Quaero, but that doesn't mean that Quaero IS Exalead. Another project partner has its own search engine (Singingfish) and the advertised main selling point of Quaero (multimedia retrieval) is not the strong point of Exalead.

So, we will have to wait a bit more until we can have a look at Quaero, but I was happy to see how capable some Quaero project partners are ... maybe it will actually work (I'm still very sceptical, but a little less than one hour ago).

Wednesday, April 05, 2006

Still Wondering about Google Base?

If you - like me - are still wondering what this Google Base thing is actually used for and if there really anyone that uses it, this post has some answers for you:

So what is Google Base? Part shopping directory, part recipe search service, part classified service.
Hitwise market share data shows that Google Base, after a fairly strong start, received approximately the same market share of visits for the week ending April 1, 2006 that it did a month after launch.
Sadly they don't have absolute numbers for Google Base usage, but they say that 0.0017% of US internet visits are to Google Base. Can't really tell if thats much or not ... but if I assume 80 mio US-Americans using the internet on any given day, on average each doing 30 visits (what exactly is a "visit" anyway?) this would come to ~40000 visits/day.

Tuesday, April 04, 2006

The Interesting Link

Windows has opened its classified site, earlier advertised as a competitor to Google Base. It is now called Windows Live Expo (you may have heard of it as Fremont). It seems to be more narrowly focused than Google Base, but better and more polished as a classified site.

About the limits of some artificial intelligence techniques: Why Data Mining Won't Stop Terror

In the Labs: Automatic Code Generators, about next generation programming languages.

Lego Mindstorms are back.

I liked this article about patents registered for Google, among them a patent that proposes "lowering the cost of wireless access by offsetting the costs via advertisements on the service", (something I predicted in January).

In more general IT related news, we saw holographic storage getting closer to market and exciting developments in cpu architecture.

More Semantic Web Clipboard

After a new post by Ray Ozzie I spend at bit more time reading the technical introduction and while I'm still a big fan of the idea, I'm sceptical about the direction the technical development is taking : Sure its cool that the Live Clipboard will work in any browser without needing any plugins, but if this means that someone who wants to enable people to copy events from her site into Outlook has to include a new css file, three JavaScript files and implement a copy callback function that returns the "Microformat" as LiveClipboardContent object - its just not going to see widespread adoption. This kind of technical realisation just negates THE reason that made microformats so successfull: the simplicity.

I know that this kind of implementation is necessary if your site wants to accept data pasted by the live clipboard - but obviously are there many more sites that only offer information than sites that accept submitted information. There should be a lighweight approach for sites that just want to publish information - and in my opinion this approach already exist: Microformats. I belive that Live clipboard needs to be integrated into a browser plugin that then identifies microformats on a page and allows to copy their structure - without requiring anything else.