Google Marker and manseo allows for this because of the unique technology behind Google Co-op. People are used to social bookmarking sites and I like them too but I was always wary of them because I felt that people can be sloppy and there was no inherent way to implement a second-tier of checking on the initial tags. Google CSEs have that built-in. It’s social bookmarking 2.0 in that sense.

Google CSE makes a great point in its documentation. Do NOT tag by keyword. Keyword searches are already handled well by Google’s search algorithms. However, search by audience (which implies intent and a level of human-only utility) is how I created the site. Also, unique sites that are difficult to pick out via keyword search are also in the list of refinements.

The first search in any Google CSE is against the entire list of URLs in the CSE’s database. The second search limits itself to a subset of the original set of URLs that are tagged by each refinement. This is not a silo because there’s no friction between the superset and subsets of data. Also, if you want to see other people’s tags on CSE, you see the tags set up by other collaborators on the CSE. My suspicion is that Google is going to put a lot of great engineers on CSE and ultimately, because of Google search technology and numerous APIs, will overtake the usefulness of old-school social bookmarking sites.

But, I personally wouldn’t use it as the *only* source to find material when researching or to keep up to date.

It is good for retrieval to find stuff you’ve come across. Although, I find that if I retag and annotate resources from the volumes of stuff I read into my social bookmarking account, that I can easily retrieve what I need via my own account.

Is that your definition of silos?

Well, if your bookmarks are in a social bookmarking system, you have access to other people’s bookmarks via the tag. I actually find a lot of valuable resources browsing by tag and particular people’s tag collections in social bookmarking services. For example, Robin Good, makes his links public. He’s topic expert on online collaboration technology. I find precisely what I need from his collection because I have access to it because we use the same service and connect via the tag.

The drawback is that not all users are identified in places like delicious. Some of the other services, like magnolia, have groups where you know the people.

I’m curious. Have you done all the tagging by point of view, “for techies” “for marketers” etc.? Right? Does it have the ability for the group to do this versus one person?

I played with some searches and one thing that struck me is that it could be a different type of silo – field that doesn’t cross disciplines.

With a Google CSE, you’re basically taking the incoming stream, reorganizing it, archiving it and then (in the future) creating yet another stream out of it. It’s a far more sophisticated feedback loop than the ones created by current social bookmarking sites. Those are JUST for tagging, there’s no way to get in there and help to reorganize it. I don’t think an nptech taxonomy is actually needed as that would be decided within the internal confines of the collaboration space of a CSE. Both the tagging AND the vocabularies are subject to a collaborative approach. With the CSE, there are NO artificial distinctions between a taxonomy and a folksonomy. It’s radically more decentralized than del.icio.us. Really. I mean it. Try it out.

The current version of Google marker actually allows nptech-like definitions at the URL level or the site level. Your argument really needs more refinement because the publisher and tagger in a CSE environment are one and the same person. It wouldn’t be hard for a tagger to create a new refinement to the tag and then start tagging sites to that site. Clearly, this suggests that Google CSE allows you to have your cake and eat it too. I think you need to really take a look at Google’s new search technologies and the APIs they’re developing. They’re going to make del.icio.us and other tagging sites look oh so 2004

Good points.

What you are describing are the pros/cons and differences between taxonomies and folksonomies, centralized versus decentralized, controlled vocabularies versus tags versus search.

I think that the mega tag stream is very broad and is like hose sometimes and many people probably don’t have the time or patient to mine it — hence the concept of an information curator or reference librarian – a human who has the content expertise to make meaning out of these distributed and aggregated resources.

I have to do a lot of web resource as part of work – developing curriculum, writing, etc – and I find that searching nptech + (other tag) in a socialbookmarking site to turn up much better and useful resources than a keyword search. However, I use both … it isn’t an either or.

and, if you were build the nptech taxonomy (which was drafted as part of techfinder at some point I seem to recall) – you’d probably start with an analysis of the tag terms used for nptech as a first step.

I think the real use of nptech is not consuming the nptech stream which is a really big content space, but consuming nptech + “relevant tag” streams.

So choosing between consuming an nptech stream and using CSE, obviously the search engine. When monitoring a specific topic nptech+cms, for example, I’ll stick with nptech.