Cscw 2010 - review of computer-science conference

Tara and I got back from a trip to Savannah, Georgia on Wed night.  The CSCW computer-science conference was there, so these are some notes from that (enjoyable) experience.

Tara had a workshop on Saturday, so I relaxed, surfed the web (reading a lot about art stuff & wetcanvas), and previewed the conference schedule. The papers weren't on a dvd this year and the usb keys were stuck in china somewhere, so pieced together what papers i could find via google scholar online. This was the first conference I was really prepared for, and it was a fun and different experience. I was pretty social, but always had a session I was looking forward to going to. Except for the two afternoon ones that we skipped to go running

Trends - lots of talks about the "facebook era," and different methods of evaluating trends in social networking. For example, I liked Ayman's poster about analysis of twitter post keywords and frequency and comparing terms with the closed-captioning and spikes with pauses in the inauguration speech.

The keynote was great - Clay Shirky was a good speaker and had a lot of thought-provoking ideas to present. One that stuck with me was the 16 year old girl (Gnarly kitty) in Thailand who blogged about the military coup but then the next post was about a photoshopped phone, and she replied to complaints with its my blog - leave me alone. To reject an audience was an interesting point. He also mentioned fanfiction.net, and its policing of plagarism (cassandra claire, when really they're all stealing premises illegally to start with). His talk went on to discuss culture, motivation, and effects. Women in wikipedia. Josh Groban charity groups and corporate identity.

work vs. Work vs. hobbies. work == hobbies to a new level? I'm somewhat confused about the difference, but I think its just in the number of hours and professionalism dedicated to it. Internally, I think, the motivations of either money or social support are somewhat similar, and its easy to treat hobbies as Work.

how is art usable? how do we interact with it? how do we learn about it? how can it be improved?

1. One of my favorite talks was the first one I attended, doing a text-based social network analysis of chat in a team group setting. Use of punctuation shows friendliness. Can determine who is manager or respected based on language analysis. The paper has some details of the algorithm. Social Language Network Analysis/Andrew Scholand & Yla Tausczik. 61% accurate using text analysis vs. 31% with just frequency of talking. Freq, duration, closeness.
   
   I imagine using it for creating an automatic friend list based on who you really interact with. Finding social power structures in decentralized communities... who follows whom, who is a leader. Who the keystones of friend communities are.
2. Adam Perer gave a good talk about social networks in IBM, and finding interesting friends based on shared stuff across networks. "Tagged with" was strongest connotator - i.e. some1 added tag 'likes cscw' to u and some1 else. Asked Q: Would you be interested to read this persons blog? After showing some matching stuff. Huge study group but limited to 500 most active stupidly. Tag_person, Forum posting, blog tags, etc. Ppl connections vs. behavior are interesting. Probs with forums for everything, so posted in device driver forum to help some1 once, or to get help.
3. Soc Tactics in Wikipedia and their effects - similar to the AP greeter culture, they discovered that 1 automated msg == 1 personal message. But more automated = negative effect, and more personal = big positive effect. Pos vs. negative feedback didn't seem to matter? Graphs showed slow decline of activity over time, which seems inconsistent with general online groups.
   
   Q's talked about "sustainability of online communities" (can't find paper?) role of newbie -> job -> leader. clear rules for the changing of rules (wikipedia admin status), path to becoming leader.
   
   Diff. behavior of newcomers to encourage feedback and interest - detecting early which members will succeed and be successful. Would let you target or de-target them, since they'll succeed anyways? Still get lost I bet from inaction.
   
   Socialization vs. work conversation - IBM-work-networks vs. true hobby/friend places.
4. Extended Families and use of webcams - jofish kaye and student. Skype kisses. Most video chats start w/phone first & upgrade to video. ~5 min frustrating setup, especially with young squirming kids. 2-3y=phone, 6-7=better?
5. "Is it really about me"Coded twitter messages - meformer vs. informer. hand-coding study, categories defined after analysis. Low reliability but some interesting data.
6. What changes since computation - a design manifesto.  Andrew Sempere @ ibm, clever, new to cscw but apparently good.    And Jonathan Grudin, who tara's always liked.
   
   Questioning the good in our shift to a digital world - a part of our culture but not all cultures.  When making new systems, question the need to replicate the old model.  Recognize users are on to you & will route around your broken ideas.  Study the gaming community.
   
   Compliance theatre - examples of routing around broken systems.  But if its working, what is he complaining about?
   
   Examples of the pre-computational culture vs. computational culture.  Attention vs. Commodity.  reputation by authority vs. practice (seems good to me, tho more complex I guess).  centralized vs. distributed control - again, more complex.  Lowest common denominator vs. customizability.  Failure as negative, vs. failure as useful study result.
   
   Q/A It's very establishment to be anti-establishment, farther out suggestions.  strongest digital controls not necessarily the best
7. Creativity & Liminality.  Staying on the edge, being pushed for uniqueness.  Diversion > Convergent.  Resilience and creativity happens when you're uncomfortable.
   
   How can cscw tools encourage this?  Boundary resources, Evoke, sustain, re-enact.  Spawn paradigm shifts vs. refining over time.
8. Computer-computer.  human-computer.  human-human.  coordination with industry.  technical and social aspects.  morphing.  repetition.  artifact driven research.
   
   I made a rant about how the open source community doesn't read conferences, papers aren't free online, and so they duplicate work constantly.  Often better and more thoroughly than we do - puts a big ? to the point of our entire research community really, if the open source community independently comes up with everything a year later.  Thought perhaps they are affected by things more than I'd guess.
9. Communities.us, teen video, fieldtrip.com  family/rural youth, internet & social media & searching.  SNS with parents, teachers ,other adults?
10. #18.  What is fun in gaming?  reminiscence in downtime of games.  Shared history / memories.  Use of "we" vs. trash talking & bond-forming.  individual actions vs. group ones.   Score page results in "i sucked" self-depreciation and breakdown of group-ness.  Q: How common is self-deprecation in games?  Insecurity?  Can it be used to judge someone's internal security feelings / self-esteem?
11.  groups in groups.  study of chinese(collectivist) vs. american(individualistic) practices and which takes precedence in groups based on majorities.  Seems pretty racist to me, but they had data to back up their country-generalizations and could tell from textual analysis which was dominant.  talkativeness vs #responses.  Chinese majority = more together / overlap w/priming.  Priming was something about introductory text before the study.  Interesting and clever methods.  What was group size?  I loved how it shows that context is a multi-level concept.
12. sonic souvenirs - interesting study with voice recordings instead of photographs with british families on vacation.  The sounds really take you back to the time in a different way.  interesting possibilities for capturing important/fun times of life with constant recording, but raises privacy and other concerns (will you use it in a fight to show who said what/etc?)  Harder to navigate afterwards.  One Q mentioned how he capture the noises his baby made while on the changing table and how it was fun to hear how it changed from 3mo-1year.
13. Moira's talk on high level autism & cmc.  aspergers syndrome and "social awkwardness".  wrongplanet.net  dating/etc. age 17-37.  lived at home, 50% drove. 10/16 lonely despite much use of social networks.  difficulty&desire of transitioning online relationship - real life.   enjoyed pre-packaged social interactions like wink/poke/like button/birthday button, but many non-austic don't enjoy these.
    
    Interesting story of not understanding that 100 text messages/day might be a bit much for someone.  How can you be intelligent but not understand that?  Difficult to imagine, perhaps a lack of empathy or just not understanding what one would like.  She talks about societal norms, but in my mind so much is relative to what I would want, not what society would.  Most people didn't have much personal email, mostly spam/newsletters (but isnt that true for all of us).  Visualization would be useful of #posts/calls/day that are normal.  Or auto-tracking these things and sending a warning message when you're outside social norms - could be funny for suggested norms @ ap too.  Similarities to workplace viz of #emails sent by each team members vs. # hours programming/LOC/metrics that help everyone feel like an equal amount of work is being done by all.
    
    Q by Gililan - austic vs. neuro-typicals - whos to say we shouldn't change.  Seems naive to me, plus their lack of understanding really is a lack of something, not an addition of something else.  Understanding your audience is always a good idea, and not being able to do so is devastating.  But perhaps not necessarily something that would be terribly difficult to automate quotas/etc for.
    
    
14. Aruna's talk about pairs of people solving a murder mystery from split-data facts using no viz (excel), shared viz, unshared viz.  ppl with split data did better with shared viz.  person with full data did well w/unshared still I think?    Has implications with police force and information sharing i.e. terrorist threats/tips/etc.  Only was 2-perosn teams, so hard to tell if it extrapolates to thousands.  Viz was only a node-link diagram, which viz community considers very primitive.  Do police use stats to decide which suspects to investigate?  race,age,subject distance from crime, alibi, arrest source, witness, have weapon, etc.
    
    
15. Next was a panel on massive data in the horizons section.  Interesting discussion about how much has changed to allow access to this data now - it existed before with phone conversations/etc but we never had access.  I should look up the facebook web app for the inauguration?  Tweet background adds a delusional quality of democracy?  We can all talk but no-one is listening?  Data vs. attention.  My q's: Trust of people in an aristotle sense, creating culture vs. changing it to see a difference.  Or just that's changing, even if we don't have an effect on it.  Attention as boredom - story of glass pieces on a handrail to force people to not use it and pay attention to the stairs.
    
    new tv = illusion of participation (realtive tv/etc).  What massive data don't we analyze?  Which data sets would have most benefit?  Subway use 4 city planning?  Healthcare use/dangers?  #credit cards used today as an economic indicator.  Convo about distopia of having search replace all science with automatic analysis of massive data.   Not really that parallel with real science at all though?

1.  

View 0 Comments & Add yours