Wherein I do my own thing
Alright, here's the deal. As those of you who were in the chatroom late on Wednesday night have already put up with me grumping about how delicious is no longer my friend. As I said starting out, I'm on Co-Op at the moment with the federal government, and my project is developing several subject guides using delicious. I have been neck-deep in tagging (on and off) for the last two months. It's not all I've been doing by any means, but it's been taking up my time in significant chunks.
Quick History & The Problem(s)_____________________________________________
There were existing guides when I started--these guides were developped with specific government groups[1] in mind. Groups have a pretty well-defined scope of interests, which makes it easier to create conceptual bundles for other people. I'm working with organic stuff, which is something most everyone is interested in for all manner of reasons. People who are looking for ways to market Canadian goods abroad want it, as do people wanting to connect value-added businesses to producers of things that can't be grown in our climate. So that's a market-related concept. There's also policy dimensions--what are the labeling & certification rules in the U.S.? In Mexico? How about in Sweden?
Nation-states and geography are a minefield. How do you tag something written by the US Department of Agriculture about the policies of the Government of Canada? And climatic/growing zones, like pollen, recognize no international borders. Not such a big deal in North America--we're stacked neatly along the North-South axis. Europe and Asia? Not so much.[2]
How relevant was all that? Impossible to know, because there was no way of knowing how big this thing was going to get. Although it became clear we couldn't give the account & password out into the wild, I'm adamant we need to get clients contributing stuff or they'll be just like every other subject guide and collect dust. I started with 15 links, grew that to 50 before trimming back.
I needed to create an information architecture based on the content that unknown people were looking for. It needed to be robust, flexible and extensible enough to handle 20, 50 or 250 links. It needed to be repeatable. So I sat down and pulled a Winnie the Pooh: think, think, think...
The Big Stuff
- What kind of stuff should I link to? [Places, not pieces]
- How much stuff should I link to? [Let's start with 25 and go from there]
Slicing Things Up
- What can you find at this link?
- What country does this cover/produced this? [can't separate without 2 sets of tags, combine together]
- What organization is responsible for this content? What organizations are topics? [again, same problem]
- What is that organization? Information from a government is not the same as information from a business organization.
- If I could sum up the nature of the thing I'm linking to, what would I call it? And how can I make that clear from the tag? [This sort of thing can tell you a lot about what you're going to find]
- I should probably set something up to make maintaining this easier. Let's make sure it ends up at the bottom, ou of the way.
The Payoff
- You should get a clear sketch of the content of a link just by skimming the tags.
- It seems flexible enough to grow a lot without breaking down.
This is what I wound up with. After the first time I restarted from scratch, anyway. I also used to tag for the official languages, which has been scrubbed.
My First Problems: Me__________________________________________________
I like to think of myself as a "ripples guy". I think a long time before I answer things, because I look for unintended consequences, like dropping a rock in a pond to see what it affects. I work on the long view. This means I overthink things--if you followed the link, you've seen exhibit A.
My way of approaching tagging and bundling is also different from many people's. I see tags as filters waiting to be applied, and bundles collect filters according to the relevant "facet", "axis", or whatever. Categories of categories.
I'm approaching this from the viewpoint that you should expand a bundle and click on a tag to get a particular variety of content. You can then use the related tags to narrow down what's available. Clicking on the bundle to bring up all the links connected to it is useless, because pretty much all the links are connected to all the bundles.
These bundles is a particular way of looking at information. Perspectives, if you like. My take is that this is probably the best way of approaching a collection of content that's all inside a particular subject area, especially if you want it to be open-ended. As the collection gets bigger, you can start creating related sub-concepts for the actual content of the links... "food" vs. "not food" (textiles, cosmetics...), for example.
Other People's Lenses__________________________________________________
Not everybody sees the world the way I do. I'm currently rebuilding my tagging and bundling schemes to go with bundles as subjects/collections of links, which is proving difficult--other people click on bundles (the name, not the arrow), and expect a bunch of links related to a topic. I put so much into creating the version 2 perspective that I'm having a hell of a time shaking it. It'll come, eventually, but I need to sit down again and do some reconceptualizing.
Congratulations________________________________________________________
You've made it to the end. Here's some cool stuff I found as a reward:
- Sign language interpretation of Gnarls Barkley's Crazy. And that's interpretation as in "artistic interpretation". Check it out (the youtube link has been killed, though he apparently has permission).
- Zooborns - Like cute overload, but without the babytalk
- The Obama Code, by George Lakoff - A *long* read by a respected linguist, which I have't gotten to yet
[1] - A "group" is a defined term in government--it's a subdivision of a larger whole. Department > ... > Section > ... > Group. I honestly don't know how deep the rabbit hole goes, but suffice it to say it's a lot like living inside the LCSH.
[2] - There's a great book about this I listened too while shelving while working at the University of Regina. I can't remember if it was A Short History of Nearly Everything or Guns, Germs and Steel. I think it was G,G&S. Both are worth checking out, if you're into this sort of stuff, though read the criticism of them too.
- murray's blog
- Login to post comments

Comments
Your post/co-op project
Your post/co-op project illustrates an aspect of social bookmarking that makes it somewhat problematic for research purposes. Searching tagged information is not yet as flexible (and perhaps effective?) as bibliographic searching via the OPAC or research databases because all tags are lumped together to define/describe the attached material. If we consider traditional cataloguing, the metadata for a particular title is divided into categories (e.g. author, title, subject) and when you search, you can search by a specific field in response to your needs. Searching for material written by author X is different from searching for material about author X. With social bookmarking, there is no differentiation between tags and what facet of the attached material they represent. So if you attach the tag "US government" to a document, does this mean the document is about the US government or is it written by/for the US government? Depending on your research needs, conducting a search using the above tag could pull up a lot of irrelevant material.
Hi Murray, You make some very
Hi Murray,
You make some very interesting points. I agree that for researching purposes tagging is not the most effective way, but by incorporating tagging within a library catalogue I think we can have the best of both worlds. This is seen in the case studies that Amanada has assigned to us. I like the idea of searching for material on a catalgue and then having a list of tags to select from to see if other relevant material comes up. Now, I realize that this is not always the case especially if someone tags an item that you or I might not see relevant to our topic area etc.
Check out my blog if you like and see what I have to say on the issue.
Lisa B
Hi Murray, I’m not entirely
Hi Murray,
I’m not entirely sure who the target audience is but are you supposed to use popular terms that other people in delicious will find, or just people you are working with? Can you have a legend/thesaurus type of thing at the beginning that explains your rationale? Then maybe you can do things like,
WTO(A) and WTO(S) for author vs subject tags (you can also spell it out in full).
Also, I do find the bulk edit feature helps a lot on delicious when you change your mind about things.