I was reading online the other day and someone said they were looking for a visualization of social bookmarking systems to show people who were totally unfamiliar with them. (I can’t find the post right now, but I’ll update this when I find the link.) I took a shot at this because it might be useful for my thesis. It’s just not simple to illustrate, so I’m not sure how effective this will be. I did this in Illustrator, so if you want to modify it here’s the .ai file.

When you view a social bookmarking system by web page you’re seeing all of the users that have saved the page to the system, as well as the tags they assigned to it. When you view a system by tag you’re seeing all of the web pages assigned by particular users to a particular tag. Tags can’t exist without web pages assigned to them. When you view a system by user you’re seeing all of web pages and tags a person has saved to the system. It’s easy to pivot between these three views, usually with one click.

What’s not shown here is that the default method of organization of items under each view is time. A secondary method is “popularity,” which differs based on which of the three views you’re in. In web page view you could see what the popular tags are for that page, you could also see how many users saved it. In tag view you could see the most popular web pages in that tag - which is calculated based on a mix of recency and overall number of users who have that page in their collection. In user view you could see which tags that person has applied the most frequently - although I wouldn’t call that popularity exactly.

The best way to explain social bookmarking is probably just to do a tour through the system, although even with that I’ve had mixed success in a presentation type of setting.

The proposition, as explained on memography.org, is simple. Create a globally unique string or meme ID. Paste this string into relevant web pages. Wait for search engines to crawl the pages. Then use the meme ID to search with 100% precision and recall. And you can even create an aboutness page to define your meme, so others can use it properly.

The above is from a very interesting post over at findability.org titled The Memetic Web which references memography.org. I didn’t have a chance to look into memography.org much directly, so what I’m responding to is the post on findability.

If I hadn’t had experience with my own name and making up globally unique strings, both by accident and on purpose, I might think it was more difficult than it actually is. To add a little evidence beyond my own experience, the people who run the site Connotea, which is for sharing academic references a la social bookmarking, wrote up a case study of their own service where they found that 14% (460) of the unique tags (3359) on Connotea were used by more than one person - that is, 86% of tags are unique strings within that system (!). Although it’s a relatively small sample I think it’s fascinating.

I use one tag on del.icio.us that I’d like to be more popular (or at least not unique), metablogosphere, which I use for sites that visualize or search in some fashion a large number of blogs. I was using this tag more when I was doing a research project on blogs, but I’ve never blogged or otherwise promoted metablogosphere. When I first started on blog research I thought it would be presumptuous to announce my ‘new’ term, and I was waiting to find the ‘real’ term. Also, how many people are actually interested in this? And would they notice and use this term? The ‘risk’ didn’t seem to outweigh the ‘reward’ - although that’s a dramatic way to put it.

Another case is aoir6 - which my class and I proposed as a tag for the Association of Internet Researchers sixth annual meeting. We used the tag for blog posts via Technorati Tags, for bookmarks in del.icio.us and for photos on Flickr. The string aoir6 didn’t exist 6 months ago (and not much beyond 3 months ago) and I did a search on Google just now and got 9,910 results, all related to the conference from what I can see.

So, in short, I can see what’s interesting about globally unique strings. My examples above have been related to tags, and I see the subject of globally unique strings and tagging as very densely intertwined.

One problem I see is that people get attached to their inventions, so if you have competing strings for a particular concept (i.e. someone has something that already means what I mean by metablogosphere) and I go out and promote my string, then you perhaps have 100% personal / sphere of influence accuracy but not a truly global type of tool. Maybe that’s not such a problem since it’s silly to think we could have a truly global agreement.

The corollary problem is that others start using a string with a different definition in mind. This is what has happened to folksonomy, which is somewhat vexing to it’s originator. This is a much more sticky problem. Does it mean what the person who originated it intended, or does it mean what most people intend when they use it? Being a bottom-up kind of person I favor the latter, but probably not if it were my invention :)

I’ve been alerted to a survey that is a pretty straightforward “Attempt to Discover what Social Bookmarking Users consider the Most Important and Desired Features in a Bookmark Manager.” The folks at blinklist.com are conducting the survey (which is more like a pole with an open ended question rather than a full survey). They’re going to share the results, so I’m looking forward to that. Any value that comes out of it will have to do with how many people actually put some thought into answering the question, which is never predictable. I’m sure anyone who reads this is quite thoughtful, so please share at:

http://www.mindvalley.com/socialbookmarking/

My answer was: Well, I love visualizations, which I usually see done by third parties with delicious data. I also love exploring to find new sites, although this process is a bit frustrating because I’m always going off on a tangent and I forget what I started off looking for. I’m generally pretty focused, but exploring these SBS I’ll start off looking for resources on social network analysis and end up on a page rating brands of swiss cheese. I’m one of those productivity people so this drives me nuts - except when my tangent takes me to something useful for what I’m working on. I think social bookmarking is going to become more mainstream but it will take some time to convince people to add one more system to their lives.

What I’m referring to as “personal data mining” here is not finding personal information about yourself or others, but looking back at things you’ve done (blogging, bookmarking, anything else) and trying to find meaning through visualization or other reconfiguration of the material.

I see this particularly with del.icio.us where several tools have appeared that help you to do this. One from a while back is extisp.icio.us - which describes itself as “extispicious, a. [L. extispicium an inspection of the innards for divination; extra the entrails + specer to look at.] Relating to the inspection of entrails for prognostication.” Other tools try to extract hierarchical groupings or provide scripts to get a list of your favorite interests (kind of a weird thing to extract when you think about it).

It reminds me of the archaeology class I had where we learned the value of trash pits for finding out how people lived. If I just think about my own garbage you could tell quite a bit about me if you took a careful look it’d be easy to find out - what I study, what kind of food I eat, junk mail, magazines, when I have a cold (like now, lots of tissues). The problem is that there’s not a good (well, easy) way for me to learn from my trash, or the paper folders I have, or the way I organize my closet–physical stuff. It doesn’t have a time stamp or metadata I can call upon to re-envision it, but many of the things I do on my computer can be re-envisioned this way.

Data mining is generally a many-to-few communication, but with personal data mining (perhaps self data mining is better - but that just sounds wrong) it’s un-reflexive self over time to analytical self at fixed points.

Since nothing is really new, I’d be interested to know what the related phenomena or antecedents to this sort of thing are?

I’m not quite caught up with all the new stuff out there on tagging, but I wanted to make a comment on the issue below from this post at plasticbag

Matt Webb and I did a fair amount of work around tagging with a project called Phonetags that I never get time to properly write up. As we were working on it, we came to realise that each of us had a radically different understanding of what a tag was. Matt’s concept was quite close to the way tagging is used in del.icio.us - with an individual the only person who could tag their stuff and with an understanding that the act of tagging was kind of an act of filing. My understanding was heavily influenced by Flickr’s approach - which I think is radically different - you can tag other people’s photos for a start, and you’re clearly challenged to tag up a photo with any words that make sense to you. It’s less of a filing model than an annotative one.

The article in general is quite interesting. My take, though, is that the file metaphor and the annotation metaphor are not so much based on the service (flickr vs delicious) but on the purposes of the user. I was just reading in a book Learner-Centered Design by Wayne Reeves that “deliberate categories [are] those designed by the individuals to meet some goal” p. 103 - and clearly tags are deliberate categories - and clearly Matt and Tom have different goals.

As anyone who has spent much time on social content web sites (flickr, delicious, etc.) knows, or takes for granted, popular does equal good. I’m not making philosophical claims but practical ones. If many people put a bookmark or photo in their personal collection it’s pretty likely that it’s useful or at least interesting. This is not quantifiable (well, perhaps, if someone wants to take that on) but it’s easy to experience for yourself on these sites.

The Daily Mashup really has something good going on. It’s just a combination of Flickr Interesting, Delicious & Furl popular and Yahoo news most clicked. A simple daily internet zeitgeist. If I didn’t have lots of things to get done I’d make it my home page.

I’d love to have a customizable version of this with lots of display options for any feed view full text, mash up with other feeds in a cloud, pick out the popular things within several tag feeds, and so on with a nice fast ajax interface. Maybe I’ll hack something together until this exists.

To clarify, I don’t think this should be anyone’s sole interface to the web leading to tyranny of the masses. Search is still incredibly valuable & someone has to put these sites on the map in the first place and when you’re talking about specific tags you can find the rare gems also.

A legitimate issue with tagging / folksonomy as it exists today is that there is no way to distinguish the level of abstraction that the tag represents (see this article for this gripe as well as many unfounded ones I might address in another post). There is no parent/child relationship among tags and so no hierarchy. Why is this a problem? I have this uncontrollable urge to see some order in my tag set and I don’t think I’m alone. When bundles came out on del.icio.us I jumped right on it. Tag clouds (like this and this) are another way to see some order but for my personal stuff a cloud was unsatisfying.

Enter RawSuger - which is, according to David Weinberger “a searchable del.icio.us with automagic, hierarchical clustering.” The impact of this automatic hierarchy will of course depend on how well it actually works - I’ve not tried it - but the possibility is exciting. Why have I not tried it? Tagging is about the community & until RawSuger gains critical mass I can’t see myself switching from del.icio.us because of a single feature, although I might try it when I have a little time.

I checked my Feedburner stats recently and it looks like there is quite a bit of interest in the new new media course–or the title was just more interesting than usual. Either way, I’ll give an update.

The tagging is a go at AoIR - it’s is even going to be in the program. Also, we may be setting up a back channel chat room during some panels. Right now the research questions for these projects are still a bit murky but we’ll work that out before October. I’m thinking the tagging project will probably be pretty focussed on the real world issues of the commons and how to grow them online, but we’ll see.

Right now I’m bracing myself for a hectic semester with working on my thesis, teaching two sections of a class, and taking this grad class. Good think I like all of this work.

Abstract: Social bookmarking systems let users store, classify and share their bookmarks online. They are global grassroots classification systems. Classification is a basic mental process that determines how we see (or ignore) the world. The first social bookmarking system was del.icio.us which came online in late 2003. I performed a pilot study of a survey of del.icio.us users focusing on feedback, motivation, and collectivity. The paper linked below is an abbreviated version of a term paper submitted for a graduate methods course in communication this past spring term. I plan to continue researching this topic for my master’s thesis over the coming year.

You can read the presentation on the Kairosnews site, or as a PDF here, or watch it as a QuickTime movie here.