Folksonomy – Classification of the masses: I don’t think that it is a way to build a taxonomy. It adds semantic information and because we share common sense we can benefit from each others tags. But it is not in a strict classification system. I use ordered tags as an advanced tagging system that does not enforce a structure but provides the advantages of complexer tagging structures — see http://tdot.blog-city.com/advanced_tagging_with_ordered_tags.htm

Compose a view of tagged bookmarks: I guess you can find a technical solution for a specific problem but in general for a users perspective I want the small structured and ordered list (Favorits) separated from the large unordered index (bookmarks database). http://tdot.blog-city.com/rethinking_bookmarks_ui.htm

Bottom line: I’m not interested in tagging as a feature or in its technical issues. I’m using it as a tool for semantic annotations. Belief it or not even without any Web 2.0 tagging system for local Firefox bookmarks (http://tdot.blog-city.com/firefox_hacks_tagging.htm)

I think your article is an important contribution to the current discussion because it questions some very frequently repeated statements, which just create a hype without understanding.

I like the way of how Rawsugar does this hierarchy thing.
Delicious has this Tagbundles but this doesn’t work as you have to manually put tags into the bundles. You offer a way to do this at the time of tagging an item..

I’m still hoping that an automated process could do this for me and all my tags I already have. But I still have to come up with a good algorithm..

Check out my page: http://rawsugar.com/links/frank and see the tag hierarchy on the left -

You should also check the way you can add a tagbox in your site with http://www.rawsugar.com/myaccount/4myWeb –
See for example http://billsaysthis.com/movies/ (top right categories and subcategories) and http://www.westernwheelers.org/main/index.htm (second half of the page)-

Thanks

Frank

I tried to add some hierarchic tags to a few links.
Is there a possibility to build a list, let’s say I tagged

• Item 1 as “internet>blogging”
• Item 2 as “internet>blogging”
• Item 3 as “internet>tagging”

When I then do a list of “internet”, is it possible to order the bookmarks
blogging

• item1
• item 2

tagging

• item 3

I just tried out and what’s cool is that when I tag something as “blogging” it is included when I search for internet because blogging is under the hierarchy of internet..

Regarding your request to put some order into your tags. Did you check what RawSugar lets you do there? You can put your tags hierarchically and thus offer your searchers (and yoursef) a neater picture of your tagspace. In addition it provides a better search. So that if you search “big firms” on your directory if will find all the links that belong to the individual firms.

Frank

I’m really curious to see what you’re working on with clusters. Did you have to screen scrape? Is there something about the API I’m not aware of?

I’m grabbing the RSS feed of recent bookmark posts each 10 minutes. That’s all. I think for building cluster data, that’s enough. I don’t need all data. I guess I’m fetching about 20% of the posts.
I also compute some statistic data.

But, this just works for building the “overall clusters”. When I want to compute which url belongs to which clusters, I have to get the rss-Feed of an url, where I’m limited to the last 30 posts. Probably that will be ok or even an advantage as people tend to tag urls diffently over time and thus it could be good to just fetch the last 30 posts of an url.

I’m against screenscraping because del.icio.us/robots.txt says:

User-agent: *
Disallow: /
Allow: /rss

And I want to respect that. I’m dependent on del.icio.us’ and yahoo’s grace..

I just wanted to address the last question, about Del’s API terms. I’m referring to the terms of use posted at http://del.icio.us/help/api/, specifically #2 and the methods available. They want you to wait a second between requests, which limits the amount of broad cross-tag correlation you can do, and they provide only methods for accessing your own tags & posts. Everything else must be accomplished via screen-scraping, which isn’t very effective on a site that changes as often as Del. The hairy math, I think, has been done: Flickr’s photo clusters are exactly what I’m imagining needs to happen, but access to the original database is a pre-requisite for that sort of thing. I guess now that they’re all under the same corporate umbrella, we may see these features drift from site to site.

I’m really curious to see what you’re working on with clusters. Did you have to screen scrape? Is there something about the API I’m not aware of?

My ultimate interest is to see an environment that continues to reward the emergence of services like Del, not so much for the features as for the incremental experimentation. It was frustrating to see the del-discuss list pounce on the author of Delirious (url?) the way they did, though he was lucky he wasn’t contending with the Y! legal team to boot.

First, I think the results of Keller’s quick survey of Google vs. Del.icio.us (using the search term cryptography) aren’t representative. Del’s audience skews heavily geekward, so it feels natural that links about math, technology and secrecy would be represented well. Picking another topic out of a hat, the results aren’t so clean-cut. Google’s results for organic farming lie at about the same level of generality as its results about cryptography. There are a lot of them, many fairly basic. Narrowing searches by adding terms (e.g. politics, how-to) helps squeeze the volume down a bit. Meanwhile, Del returns just 10 links, total. Hope they’re good. Conversely, the quality of results for terms like web or blog in Del.icio.us is also generic and fairly undifferentiated. Google maintains a middle-distance from the user in more situations, while Del.icio.us swings wildly based on the composition of its user base.

Hmm.. probably you are right. I should take out google search from my “distance to data” image. On the other hand: Google search just gives me results to my query. What I wanted to analyze was, how does the web service help me to navigate, how do I see related terms, etc. Google doesn’t give me these infos (sometimes it helps me a little, for instance when searching for internet, a news entry about internet shows up.). What Google does is ordering the items to hopefully show me the more relevant articles to my search term. This is a feature lacking in del.icio.us and DMOZ. Ah, maybe I’m comparing apples and oranges..? Maybe we should view browsing in tagged content as an alternative to “browsing” in search engine ranked data?

Second, I’m losing faith in the potential for staying informed from tagged web content alone. I’ve been subscribed to several Del tag searches via RSS (e.g. attention, maps, infographics), and today I unsubscribed from the last of them. The data was far too general: I was either getting a very slow trickle of information, or a torrent. Attention in particular picked up sharply from fewer than a half-dozen items per day a few months ago, and accounted for a substantial percentage of my RSS volume. Many of these items were repeats or pointers to the Attention gang’s own writings. Interesting, but I get this stuff via other channels, and wasn’t learning enough new stuff.

I second you. That’s exactly what I feel too. I subscribed to del.icio.us popular on “tagging” and “journalism”. Both give me some good links. Just 1-2 per day, and just the popular ones (I too have no faith into the popular thing in del.icio.us anymore). That’s why I urge to have better possibilities. I want clusters!

The quality of the links was highest when the system was populated by hardcore early adopters, and dropped when middle of the bell curve moved in. In particular, the quality drop was characterized by echoes and repetitiveness. I would argue that Folksonomy Helps Me Stay Informed On a Given Topic best when the diversity of a given tag/user/resource population is high and the headcount low.

Quote of the day. Nothing to add here.

I expect that Del.icio.us could do something like this too, but the API terms are too limiting for external experimentation, and the math is hairy. If anyone had the gall to do this on a large scale, it’d be Google – they’re already treating link text as a de-facto tag on a page, so they’re arguably the oldest tagging outfit on the block.

Can you explain what you mean by “but the API terms are too limiting for external experimentation”? About the hairy math: Nobody has done it but I believe it is possible. I’m doing some hairy math on del.icio.us tagging data (I’m computing tag-clusters) and the results seem very promising so far. I hope I have a service up in a few months.