March, 2008


28
Mar 08

Fixing Information Overload in Twitter

As someone who has started or run a few web projects, I’m used to the complaining blogger. And because of that, I try to stay away from being the complaining blogger. But I think that Twitter is about to drive me crazy with information overload, and I think I know how to solve the problem. So here’s a go.

Increasingly, Twitter has begin to feel like a collection of RSS feeds. My Twitter home screen is my personal newsreader. Unfortunately, it is completely dysfunctional. Some people I follow for personal reasons, some people I follow for work. Some people post often, some people once a week. I want to read every single message written by some people, and others can float by. Sort of like any other inbox, I guess.

If you’ve used Twitter, however, you know that all your messages go into the same place. Everyone is treated equal. There’s no method to deal with the information overload inherent in the system, there’s no way to mute over-Tweeters, there’s no way to have any control over the information space.

This may have worked in the early days of Twitter, where the interaction was supposed to be ephemeral – some messages you caught, some float by, who cares. Unfortunately, Twitter has grown up as it became more mainstream. People are saying “Did you see my Tweet” just as they would say “Did you get my email.”

We know how to deal with this: Folders, labels, mute buttons, regular expressions, etc. We need Tweetboxes, we need Tweetfolders to separate contexts, we need better strategies to deal with the information overload. And this is just in regards to incoming information – the multiple-audiences problem is another, more difficult problem.

Looking around at Twitter clients, I don’t see any that support such functionality. But I’m not really interested in using a Twitter client – I just want Twitter’s web interface to work. Let me create some folders, or tag my contact into a few different bins, so I could sort my incoming messages. A mute button would be nice as well, but right now folders (or labels, if you want to be Gmail-y) would really help. Look at the existing patterns that work with RSS and inboxes, and give us that. Because this current all-or-nothing isn’t the right answer.


28
Mar 08

The Social Filter

Over on techPresident, I’ve written a post about the social filter, in response to an article in the NYT that caught my attention. My post is about how we’ve created a new media in the spaces between us in social media. It was fun to think about and write – check it out here.


25
Mar 08

The Best Social Software

Over the past few days, I’ve been discussing the problems of multiple contexts in social software. While I am primarily covering this problem in the context of Facebook, this is a problem that affects all social software. I’m actually feeling this problem most acutely with Twitter these days.

When we use social software, we adopt a persona. Broadly speaking, that persona is mediated by the audience and publicness of the space. For example, this blog is hyper-public (Googlable) and the audience is mostly research- and tech-folk. As a result, I write on topic and try to stay away from stuff that is going to embarrass me too much down the road. A simple equation, but I think it works.

At some level, we all engage in this audience/publicness calculation when we craft our persona in social software. Blogs charted this territory; the central tenet of “successful blogging” is knowing one’s audience. But blogs are usually hyper-public, meaning the nuance of publicness is lost. Let’s consider private Twittering as an alternative.

When one elects privacy measures in Twitter, they limit their publicness. That is, they exercise control over who sees the persona they are creating, a persona that is a function of audience. As private Twitterers construct these trusted places – they understand their audience/publicness calculation – the persona becomes more personal, the content more engaging. Remarked to me via a private Twitter message, “it feels like we have insight into ppls thoughts thru-out the day.

The best social software should make you feel like you’re amongst friends, encouraging you to create a more true persona. The best social software lets you be you – whatever that you happens to be at the moment. Facebook 2005 hit that sweet spot, and Twitter affords the more geeky of us that place today. Perhaps this is why every time I log into Facebook all I see are Twitter status messages; vibrancy lies in the more personal network.

As contexts collide, however, the audience/publicness calculation has to be reworked. I don’t dare look at the hundreds of Twitter follower requests waiting for me (most spam – I’m not that internet famous yet), because I know allowing more people in to my circle would force me to refactor myself in Twitter. And I don’t want to do that yet, not while the experience is still so great. Ultimately, though, I’ll have to, and that will be the death of Twitter for me. The publicness will force depersonalization, and my Twitter will become like my blog.

Is this process unavoidable? I’m not sure. But the expectation of hyper-publicness ultimately written into social software needs to be rethought. As we force users to constantly renegotaite the audience/publicness calculation, I think we lose more than we gain. Rather than ultimately forcing publicness, we need to think of ways for users to create private spaces for sharing. This is why so many people love and care about LiveJournal to this day – it allowed the creation of private spaces.

Most of us are not internet celebrities, but the social software we use assumes we are (or want to be). It’s time to rethink this, to build closets and spaces for whispering into social software.


24
Mar 08

New Course: Relational Perspectives on CMC

Sometime last semester, when I had obviously lost my better senses, I agreed to create a new course for Fall 2008. Well, the course is now on the books, so I thought I might post the description here:

INLS 490-151, Fall 2008
Computer Mediated Communication

The web is a place of communication, interaction and relational management. From Ray Tomlinson’s first email in 1971, to the iChats, Wall posts and Twitters of today, we have consistently turned to the network to find one other for communication and collaboration. This course explores computer mediated communication (CMC) from a relational perspective; that is, how do we use network communication tools to start, build or sustain interpersonal relationships. Utilizing interdisciplinary perspectives including, but not limited to, information science, communications and science and technology studies, we will explore the theoretical, practical and historical perspectives on the role networked communications tools play in the relational process. Students who successfully complete this course will develop critical perspectives on networked communications, better preparing them to develop the communication tools of the future. The course will be structured as a seminar, with substantial reading and discussion. In addition to presentations, students will be expected to develop hands-on experience with the communication tools of the past, present and future.

As the abstract describes, this course is about how we use networked communication tools to manage and sustain relationships (of all sorts). The interesting turn in this class may be the inclusion of a historical perspective – the network communication tools space is so fast moving, that I think mastery of critical historical perspectives will provide a strong basis for critique of future tools.

As I develop the syllabus over the summer, I’ll be posting updates. If you’ve got any syllabi you think I should look at, tag them for me in del.icio.us or post them in the comments. I’m developing a nice archive of syllabi here, so I’m always looking for inspiration. This also means I won’t be teaching my online social networks course in the Fall, but I think I may teach it once more before I graduate.


21
Mar 08

The Perfect Virtual Community

In yesterday’s post about Facebook’s new privacy system, I discussed the concept of “community health” in online social networks. This is a topic I’ve thought about for some time, and explored in my essay The Vibrancy of Online Social Space. What is a healthy, vibrant online social network? How does one build or shape a social network (or other virtual community) so that it is healthy and vigorous, an approximation of our best cities or communities?

This is actually a very important point – one that I encourage social entrepreneurs and community managers to ponder; it’s never enough to just throw affordances or rules at a community, a community must be gardened with love.

Remembering Facebook ca. 2005 (or even Friendster ca. 2003), we can reflect on how the community has changed. In yesterday’s post I talked about “privacy” as a key proxy for gauging community health. In early 2005, everyone in Facebook felt like they knew one another; your audience was your network, and your network was your friends (or potential friends). As a result, we didn’t use privacy, we disclosed a lot, and we engaged each other digitally at a level never before seen.

At the time, when I began studying the community, I sensed there was a privacy divide, that young people don’t understand or care about privacy like “we” do. Over time I’ve realized I couldn’t be further from the truth. To those users, Facebook in 2005 was the perfect community, a digital place they felt so comfortable with that privacy didn’t enter the equation. It would have been as weird to use privacy in Facebook ca. 2005 as it would be to walk around with a bag over your head on campus today.

And just think about that for a minute – the perfect virtual community. That’s a remarkable achievement, and much credit to Facebook for creating such a remarkable success. Unfortunately, as Facebook opened the doors widely, they learned that community doesn’t scale. This isn’t new – danah boyd documents the clash of communities in Friendster in her paper “None of this is real“. As contexts collide and communities become more heterogeneous, virtual communities become more real – and the privacy fears and stranger-danger that come with real-world networks erode our feelings of community and cohesion.

The Facebook of today is vastly different from the Facebook of 2005. With the influx of new people and new networks comes the clash of contexts. This forces us to put locks on our doors, to shut ourselves off to all but our friends, to confront the non-idyllic parts of community.

Reflecting on Facebook 2005 and Facebook 2008, I think there are important lessons to be learned – for makers of social software, for community gardeners, for those who might wish to make a living at this one day. What can we learn from Facebook, and how can it be applied to the communities we’ll construct tomorrow? And can we ever have a community as strong and vibrant as Facebook 2005 again? I certainly hope so.


19
Mar 08

Facebook’s New Privacy Settings: Too little, too late

This morning, Facebook introduced some fairly significant updates to their privacy controls. Documented in this Facebook blog post, the changes are:

  • Facebook has rolled out a consistent privacy interface, which allows access to shared elements based on access-control lists (i.e. work network, school network)
  • These access-control lists (ACL’s) have been expanded to include ad-hoc groups of your creation. Therefore, it’s possible for you to share some elements with only your work friends, and others only with family, etc.
  • Finally, Facebook has changed their network-based control model to allow friend-of-friend access. That is, you may now share things with your friends of friends that aren’t in existing networks. This is a big departure from Facebook’s operating plan to-date.

I want to begin by giving Facebook a lot of credit for the standardization move. As an outsider looking in, I’ve always sensed a HCI/UI-vs-BizDev disconnect when it comes to privacy. Facebook actually has very elegant and granular privacy controls, used most extensively by power users, but they’ve always been there. This attention to detail (the engineering and UI challenges of deploying item-level privacy are not trivial) always clashed with ham-fisted efforts like Beacon or privacy-less Newsfeed. Score one for the engineering team for the development of the consistent privacy interface, which is a good move.

Now let’s consider the business implications of these changes to Facebook’s privacy model. Facebook is trying to solve two problems here – the context problem and declining core-user pageviews. With regards to context, Facebook’s users are facing the problem of multiple contexts: what happens when my friends, my boss and my parents are all my Facebook friends. As Facebook becomes less about our everyday friends and more about our bosses and coworkers (or people you have to sit across from on Thanksgiving), Facebook naturally becomes less interesting, with people sharing less. It’s hard to manage these jumbled contexts, to know who you should and shouldn’t be disclosing to, especially when one has 500 or 1000 friends.

With context jumbling comes a natural move towards privacy. As Facebook has expanded, its cores users have increase privacy and shut their profiles off from the world. Gone are the days of wide-open Facebook; in a recent pilot survey of Facebook users (average age 25), 86% reported they use privacy settings in Facebook. Why? As more users have joined, as contexts have jumbled, Facebook has transitioned from a friendly community where no one kept locks on doors, to a normal, mundane community where one locks the door and shuts out strangers. Remembering the Facebook of 2005, this place where everyone shared with one another, one can’t help but wonder just what Facebook lost as it forced users to confront the real world via Facebook.

With the addition of contact lists, Facebook is taking a stab at solving the context problem. Theoretically, one can segregate one’s friends, family, best friends, roomates, and so on into private networks for selective sharing. Of course, when you have 500 contacts, it becomes rather difficult to remember who belongs where, or what lists contains what friends/family. Contact lists are bubblegum in the dam when it comes to the context problem; it will prove useful to some, but most hardcore users have such large networks that the contact-management process will be challenging. I expect most users to create one, maybe two groups. Of course, if they get value from that, it’s a win for Facebook.

By adding friend-of-friend optional sharing, Facebook is trying to address the smothering privacy trend moving through the system. In our pilot study, 88% of users reported viewing less than ten profiles a day, with 35% of users viewing less than three profiles per day. As privacy has increased, the value one gets from the browsing process has decreased. Have you tried to browse anyone’s friends recently? It seems that all you run into is private profiles. By allowing friend-of-friend connections, Facebook hopes to make browsing a popular function again, one that increases ad and page views. Newsfeed, cluttered with spam, has become less useful for generating pageviews – so Facebook is turning back to what made the service so initially valuable – our interest in one another.

I hate to say it, but this is a too-little, too-late move on Facebook’s part. Privacy is epidemic in the community, spurred by media narratives and self-regulation. Unlike Beacon or Newsfeed, these changes are an opt-in measure, meaning that only intentful users will switch their privacy settings. Unless Facebook figures out a neat gimmick to get people to buy in, they will have a challenge in pushing adoption.

Stepping back from this initiative, I think there’s a valuable lesson here for others managing virtual communities. Its much harder to ad-hoc technical fixes onto jumbled communities after the fact. It is also extremely hard to scale community effectively; Facebook’s initial segmentation allowed expansion without problems for some time, but ultimately, as the friend requests from the uncles and old friends you’ve never seen in ages pile up, the place became one where any rational person would be afraid to “live publicly.” Unfortunately, this cat is out of the bag for many of Facebook’s users, and I doubt that friend lists will solve the problem.

What do you think?

On an unrelated note, why does Facebook’s blog have a comment form if it doesn’t allow comments?


19
Mar 08

Practical Unit Structures: BibDesk Importing and Templates

Over the past few weeks, I’ve been organizing my citations in preparation for proposal writing. I’m really terrible at the citation-management process; I just let pdf’s pile up in folders on my desktop, putting them into a citation manager every few months. The reason I put this off is the process is so tedious – hand-entering citations is about the last thing I ever want to do.

Today I’m going to share some tips that may make your citation-management process less painful. If you’re a grizzled academic you probably know all this stuff, but fellow grad students may benefit from this. First and foremost, I use BibDesk as my citation manager. BibDesk is for the Mac, it is free, and it is one of the best pieces of academic software I’ve ever encountered. I highly recommend it if you’re looking for a citation manager.

The first bit of advice is with regards to the citation-import process. This is fairly general advice and not specific to BibDesk. If you’re tired of hand-entering citations, many times publishers will provide downloadable citations from their websites. For example, if one of your papers is hosted by Sage, Sage’s site will allow you to download a citation directly into your citation manager. Generally these sites will provide the citations in various formats; the good news is that BibDesk will import almost anything. As an example, here’s a link to a paper. To download the citation, simply click on the “Download to citation manager” button and you’re set.

Different publishers have different approaches and techniques, so you’ll have to sift through the sites to find the option for downloadable citations. ScienceDirect is particularly stupid, requiring you to be logged in to download citations. Other sites spit out invalid files, so you’ll have to touch some stuff up by hand. Its somewhat incredible to think that publishers haven’t got citation exporting right, but I’ve learned not to be surprised by academic publishing.

Now, if you can’t find citation exporting on the publishers site, the academic search engines Google Scholar and Microsoft’s Live Academic provide downloadable citations. To turn on citation downloads in Google scholar, open the preferences and select “Show links to import citations into [your bibliography manager].” A link will appear next to all search results allowing you to export the citation directly into your citation manager. Beware, however – the results are often incomplete, incorrect or referring outdated versions of the paper. However, I’ve saved more time than I’ve lost with Google Scholar, so it might be useful to you as well.

The previous tips will work with any citation manager; my next tips are BibDesk-specific. BibDesk is built for integration with LaTeX, which works fine for some of my papers (but not all). Sometimes I just need to export a list of APA citations which I tack onto the end of Word file. Unfortunately BibDesk makes this difficult by default. To solve this problem, you need to use Templates.

With export templates, you can define a custom format for exporting citations that can be included in a document. Unfortunately, there’s no built-in for exporting to APA (or any other format that might be useful). I was able to find some APA-like examples, but none that fit APA exactly. So I went and hacked on a template, and now I’ll share it here. Linked here are two templates, that you can import into BibDesk. The first is APAInlineTemplate.txt, which will generate the (Lastname, year) inline citations you can include in your document. The second is APAFullCiteTemplate.txt, which will generate full citations to include at the end of the document. Download them both as a zip file here.

To install these in BibDesk, open up Preferences->Templates, and then click the plus button. To install, give them names (“APA Inline Citation” and “APA Full Citation” work for me), then select the files (you’ll first want to move them to ~/Library/Application Support/BibDesk/Templates). Use txt as the type. Once you’ve got them installed, right click on citation in BibDesk, select the “Copy Using Template” dialogue, and then select an APA style. You’ll now be able to copy APA-standard citations directly out of BibDesk for inclusion into your document.

I should also note that I am only responsible for hacking these templates – I believe I downloaded the originals from the BibDesk site and modified them. I’ve added conditionals to make them more forgiving (say, you don’t have the Address of the publisher) but there may be weirdness if you’ve got funky data. And the template doesn’t address every style of publication – but it covered most of what I had in my library. To find out more about templates, read the BibDesk wiki page here, and here’s a link to the manual.

This stuff is pretty dense – it took me a while to figure it out, but I’m glad I spent the time. BibDesk is now much more useful to me, and I hope I’ve saved some time for you as well.