Gladly, there were also more relevant projects, i.e. studying the usage of RFID technology in hospitals and the design of landmine seeking devices (and helping people to do this safer). Personally, I would have preferred more proejcts like these, along the lines of the OLPC project.
From a technical perspective, several proejcts I found notable: Sensing the direction of wireless signals through the use of spinning antennas on routers using the Doppler effect (the effect that can be heard in the siren sound when being passed by police/ambulance car) and a technique to sense the walking direction of a mobile phone user based on the construction of the human hip bone. Also, a notable study researched how much power can be gained from everyday human movements by theoretically adding small acceleration-driven dynamos (< 10g of weight, if I remember correctly) to the arms, ankles, etc - I find this a little bit scary, actually. (Just as a side note: The reportedly possible power was very low, but reportedly enough to power a MP3 decoder or GPS receiver chip.)
The workshop I attended was "Ambient Information Systems 2008", concerned with mostly visual ambient displays, and also a tactile one (I presented the 'Ambient Life' project - a physically breathing and pulsating mobile phone). With 15 people attending, we were a comfortably sized group, and after everybody had shown his work, we began to discuss the 'grand/meta issues' of ambient displays in the next years. (More information about this can be found at www.ambientinformation.org)
Further topics of UbiComp 2008 were context awareness, wearable computing and 'always online' - of which the latter I found particularly interesting. A study presented was concerned with the interruptions and problems concerned with mobile IM; users on both sides of the communication line were observed. Interestingly, most who started a conversation ignored the availability status of the other, and every fifth user started with a sentence that was relativized by including the word "just".
The conference itself is a very interesting attend, which I can recommend: Especially because I think designers could add much and gain a lot from the dialogue with researchers of this field.
I give UbiComp 2008 a rating of 4/5 stars.