Why you should not trust Sheridan Printing with your conference paper

In 2009 I found a pretty obvious security flaw in Sheridan Printing’s submission management system. It allows anyone to view and modify all papers in the conference proceedings of many major computer science conferences prior to printing and publication.

Over the last two years I have continuously tried to get this problem fixed silently – without success. Therefore, I publish the issue now, giving authors the chance to make informed decisions.

In this blog post I describe the problem, explain its possible consequences, and propose ways to fix this issue.
Read the full post »

Nur ein Test, ob Google ↻ findet.

Guttenberg und Rieble

Wollte gerade die Dissertation von Herrn zu Guttenberg aus der Unibibliothek ausleihen, ist aber schon weg.
Dafür hier ein Suchbild:

Guttenberg und Zehnpfennig


Gerade hat dpa gemeldet, dass es wohl Anzeichen gibt, dass unser Verteidigungsminister evtl. in seiner Dissertation abgeschrieben habe. Also gleich mal online nach der Dissertation gesucht und gefunden:


Die Einleitung beginnt mit

“E pluribus unum”, “Aus vielem eines” – so lautete das Motto, unter dem vor rund 200 Jahren die amerikanischen Staaten zur Union zusammenfanden, und dieses Motto ist programmatisch zu verstehen. Das Land, das wie kein anderes den Pluralismus auf seine Fahnen geschrieben hat, eröffnet erst auf dieser einheitlichen, gemeinsamen Basis den Spielraum für die Entfaltung von Vielheit.

So schön geht es seitenweise weiter. Wenn man nun ein paar Satzfragmente bei Google sucht:

Read the full post »

Kernel talking to me

Message from syslogd@seamless at Nov 11 04:01:37 …
kernel:[2972477.277368] Uhhuh. NMI received for unknown reason a0 on CPU 0.

Message from syslogd@seamless at Nov 11 04:01:37 …
kernel:[2972477.277368] You have some hardware problem, likely on the PCI bus.

Message from syslogd@seamless at Nov 11 04:01:37 …
kernel:[2972477.277368] Dazed and confused, but trying to continue

Message from syslogd@seamless at Nov 11 04:01:37 …
kernel:[2972477.277368] Uhhuh. NMI received for unknown reason 34 on CPU 0.

Message from syslogd@seamless at Nov 11 04:01:37 …
kernel:[2972477.277368] Do you have a strange power saving mode enabled?

Message from syslogd@seamless at Nov 11 04:01:37 …
kernel:[2972477.277368] Dazed and confused, but trying to continue

Best {Paper, Demo, Poster} Awards Considered Harmful

Many academic conferences award one or more “best paper”, “best demo” or “best poster” awards. The awardees are either selected by the program committee or by an anonymous audience vote.
However, in my opinion, we should get rid of these awards.
For three reasons I think these are a bad idea.

1. The big question regarding such awards is: “In which way are they useful for the community”. Like in education, an intervention (the award) should have a lasting positive effect either on the awardee or the community. Sure, the individual author who receives an award might feel happy for a short while. However, this positive effect might also be reached by just patting them on the back and saying “Great Work!”. There is no evidence that such awards lead to higher achievements. Quite to the contrary, a number of publications claim that awards and incentives actually have overall negative effects on individuals and communities, lowering their performance [1].

2. Awards are a bad metric for great research. Bartneck and Hu have pointed out that (on average) papers which got a CHI best paper awards did not get more citations than a random sample [2]. It seems that even a commitee of experts is unable to predict which papers will have the highest impact. If detecting great work is not even possible for papers with their fixed structure, why should it work for posters or demos? Especially demos are so diverse that a one-size-fits-all award is plain wrong. Is an artful, thought-provoking demo better than a demo of a novel, extremely versatile sensing technology?

3. Undersampling is another problem. How many of the conference attendees have seen all posters and demos? I would guess that, for any poster session, not a single attendee has read all poster titles. Likewise, it is hard to judge the quality of a demo without understanding it. For understanding a demo, you have to try it out for some time. With hundreds of other attendees trying out the same demos, there is just no time for this. Therefore, almost all votes for a best demo or poster can consider only a small subset. Nearly noone is able to make a qualified decision. Variables like poster/demo placement or group dynamics might have more of an effect on the votes than any kind of actual “quality” of a poster or demo.

Overall, best poster/paper/demo awards are neither shown effective nor valid nor at least fair. Why are we then clinging to them?

The TEI conference – which has a very intensive and diverse demo and poster session – has opted not to have any awards – for more or less these reasons.

[1] Kohn, A. (1993). Why incentive plans cannot work. Harvard Business Review, September 1993

[2] Bartneck, C., & Hu, J. (2009). Scientometric Analysis Of The CHI Proceedings. Proceedings of the Conference on Human Factors in Computing Systems (CHI2009), Boston, pp.699-708

ITS 2010 – Day 2

I’m attending ITS 2010 – the ACM International Conference on Interactive Tabletops and Surfaces 2010 in Saarbrücken, Germany. This is a short collection of interesting stuff I’ve seen and heard on day 2 (Monday, 6. November 2010).

(The demo and poster session is like a huge, dark playground with (literally) tons of amazing touch interfaces.)

Monday was the first day of paper presentations. There was a wealth of papers on several topics. Therefore my account is very selective. You can get all papers at the conference website.

The day started very relaxing with “Tafelmusik“, two musicians with a digital audio sequencer and a table full of objects that make sounds. See their website for a photo. By sampling them and continuously replaying these sounds they created a sound landscape – sometimes soothing and sometimes fascinating.

Brad Paley gave a keynote covering a wide range of topics but centering about ways to visualize information. Some of his claims:

  • “CHI” considered harmful: instead “computer mediated human-to-human interaction”
  • Color is bad for encoding data
  • Consistence *impairs* performance
  • 15:1 increases in information density, 20:1 speed-ups can be easily reached
  • “users” considered harmful

While Brad did not explicitly say so, I think in their entirety these claims only apply to UIs for expert users, however.

Afterwards, Malte Weiss (RWTH Aachen) presented “BendDesk: Dragging Across the Curve” [PDF]. He and Simon Voelker built a desk with an interactive surface bent partly upwards. Malte kindly mentioned Curve – our research on this topic. We are currently figuring out how to connect both prototypes for remote interaction.

In the same session, Yvonne Jansen presented MudPad [PDF], a tactile display using ferrofluid and magnets.

Antti Virolainen presented an interactive surface made out of ice (FTIR in ice is probably not possible).

Hrvoje Benko (Microsoft Research) presented another spherical multitouch surface – but this time a large dome where you walk inside [PDF]. Interesting link from his talk: worldwidetelescope.org

In the afternoon, Dietrich Kammer (TU Dresden) presented an interesting theoretical framework for describing gestures [PDF].

For me, the demo and poster session is always the highlight of a conference. At ITS 2010 it took place at DFKI. There was a wealth of really cool demos and interesting posters. As I had to present my own poster (“Some Thoughts on a Model of Touch-Sensitive Surfaces” [PDF]), I did not find time to have a look at every demo. However, there was an amazing mixture of art, high-tech hardware, and applications. See the photos on Facebook!

While I liked some demos and posters more than others, I did not fill out my voting sheet for best poster or demo. More on this later.

Photo taken from the official ITS 2010 Facebook album:

ITS 2010 – Day 1

I’m attending ITS 2010 – the ACM International Conference on Interactive Tabletops and Surfaces 2010 in Saarbrücken, Germany. This is a short collection of interesting stuff I’ve seen and heard on day 1 (Sunday, 6. November 2010).

(people testing the multitouch device they just built)

Sunday started with four tutorials:

Sheelagh Carpendale (University of Calgary) gave an introduction into qualitative evaluation and observation of tabletop interaction. We did an experiment where she would tell us a story and we should draw a continuous line somehow representing the mood of the characters in the story. I found it quite interesting that there were completely different looking drawings but some drawings looked quite similar. I wonder whether you could find out which persons have a similar understanding of a story. Looks like a great method for match-making.

Uli von Zadow (Archimedes Solutions) gave a very interesting overview of multitouch sensing APIs and implementation details of common processing stages (e.g. it is not a good idea to put successive processing stages on different processors as this makes the processor cache pretty much useless).

Florian Echtler (Hochschule München) presented an overview of multitouch sensing techniques [I did not attend the talk as I’m pretty familiar with his work].

And finally, Anne Roudaut (HPI Potsdam) organized a really cool “build your own multitouch” session where about 50 participants built a simple touch-sensitive surface using FTIR with visible light and a cheap webcam. You can find the instructions online on her Acrylicpad page. They got really cheap (5 EUR) webcams on eBay. However, they only worked with Windows XP – and of course Linux 🙂

Afterwards we had a nice get-together at the Ratskeller in Saarbrücken.

Photo taken from the official Facebook album:

Day 1, Day 2

Emotiv EPOC Brain-Computer Interface cracked open

The Emotiv EPOC is a relatively cheap ($300) EEG headset intended for gaming.
I have considered buying such a device for some time. Unfortunately, the device is sold out at the moment and only available in the US.
A developer headset ($500, sold out, too) is available worldwide. However, you would need to purchase an SDK, ranging from $500-$7500 in order to do anything meaningful with it. The SDK does not support Linux, of course.

However, today Cody Brocious (Daeken) has released the first version of a Python toolkit for the EPOC. For this he cracked the AES encryption and partially reverse-engineered the protocol.
There’s still a lot missing – for example, it is not clear which signal belongs to which sensor, and the signals are not yet analyzed in any way.

This seems like the right opportunity to get an EPOC and tinker with it. Thanks Daeken.

Some discussion is happening at Hacker News and Hack A Day

Update: H+ magazine has an interview with Cody, conducted by Bryan Bishop who already wrote a Python wrapper for the EPOC in June, using the official SDK, however.

Update 2: It seems the developer headset is different from the standard one. skadge is trying to get it working.

Update 3:
Emotiv responds in their forum:

We strongly oppose this kind of action. It violates the end user license and seriously threatens our viability. It is no exaggeration to say we may be forced to withdraw the consumer headset from the market or go out of business entirely. Kinda defeats the stated purpose of “opening up” the product for research users – which is frankly spurious because it IS open to properly licensed owners – and OpenViBE has a properly licensed driver already in Beta development.

It’s all well and good to demonstrate how smart you are by hacking the iPhone – Apple can afford to lose some revenue. We are a struggling company and this action has prompted our investors ask us to show cause why they should not demand repayment of our operating loan. Research licenses are a substantial part of our revenue. Thanks to this selfish action we may well have to close the doors or withdraw the consumer product, or re-price it at a significantly higher level.

That enough discussion for you?

Update 4:
Daeken has posted an unconfirmed Dev Key on IRC:

conKey = '\x31\x00\x35\x54\x38\x10\x37\x42\x31\x00\x35\x48\x38\x00\x37\x50'
devKey = '\x31\x00\x35\x48\x31\x00\x35\x54\x38\x10\x37\x42\x38\x00\x37\x50'

Computer Science Conference Domains Grabbed

Seems like someone grabbed a number of domains for upcoming computer science conferences like ITS 2011 and 2012. The (anonymous) registrations took place on July, 23rd 2010.

I’m curious whether any conference organizers will indeed buy back the domains.

The URLs registered so far are:
Read the full post »