Library clips

This is more a wish than a prediction but I am betting that there will be one killer search engine for all the social bookmark services, someone to do for bookmarks what Technorati does for blogs.

So far we have a lot of meta-search tools:
Gada.be
GutenTag
Tag Central
Tagbert

Wink is the only tool that is merging the results but how are they ranked is the question.

What you’d think happens is that the item that has been bookmarked the most in all services with that tag will get top ranking and so on…I guess this is TagRank.
They could also mix in a Google link search and Technorati Link search to the tweak the ranking I guess, but the idea of just pure Tag Rank is good I suppose.

For each result set a link to the tag in all the services would be good, and as Wink already does, is listed related tags.
What about common tags for each item in the results from each service…this is extending the search to discovery…and/or common tags for the result set.

Since this is a human indexed web, what about being able to search full-text, the relevancy and ranking would be something similar to Google’s Page Rank with some of the TagRank thrown in

…the results would suggest looking in various tags, or see results in clusters.

This way we could have a human indexed web (less noise) that is searchable by full-text (relevancy/ranking a la PageRank and TagRank), or search via tag (relevancy/ranking according to TagRank)…like subject searching on an OPAC, as well as full-text searching.

So far the closest service has been Zniff, but this is just for the Spurl service…if all the services were combined this would be a powerful and cleaner search engine, that enables to limit search to a subject category.

Ohh, I forgot, there also has to be full-text search within a tag
…these refined results could again be based on a type of PageRank.

Speaking of refining, when you search for a tag, a list of intersecting tag could be available to refine your results, also boolean tag searching would be great…just check out Simpy for great searching features.

So who’s it gonna get into this risky game.

Imagine doing a link search on your blog domain to see which posts and how many times they have been bookmarked by the various services, and what tags people used to describe/file your various posts.

Here are some of my earlier posts:
What is the tagosphere?
Folksonomy: search vs. browse
Social bookmarks vs. free text search
Tagging alone is not a panacea for retrieval!

So the next time you search wouldn’t it be great to be able to choose the result set based on PageRank or TagRank.

[ADDED 22/02/06: rel8r]

Technorati have added new features…the search results now have a user link, clicking on this will take you to the users Technorati Profile, from here you can click on a tag from the tag cloud where you will see posts from one tag from the one blog.

I wonder if there is a way to limit a search in Technorati Tags to one blog (I mean it’s all there now)…IceRocket is on the game.

What about a link to the whole tag cloud…and some code for your sidebar…

Also, every item in the results page has 2 icons: one to see incoming links for the post, the other to refine the search to just that blog…I’ve got to say Technorati is looking super of late, and is super fast…good on ya guys!

Actually there’s more, depending on your search different labels appear in a tabbed format, so if you search for Gadgets, this of course is searching in all Blogs, but clicking on a tab will refine the search to blogs about: Technology, Apple, Tech, IPod, Hardware, Music…this is a great use for the Technorati Blog Finder data.

They are also handy widgets on the sidebar, one is keyword trends…DataMining has more.

Phil Bradley’s I want to guide has been extended, there is also a blog to keep up with the latest.

There are so many ways to make these type of lists: webpage, wiki, blog, bookmarks, outline…for me this seems like a type of classic link blog.

The quickest way would be to bookmark each resource in a social bookmark manager, add a comment, and have that RSS feed re-syndicated to a blog
…or even just be content with the bookmark manager as the portal, but I suppose the blog format is going to be more noticable.

Reminds me how Emily Chang posts.

Memeorandum (smaller scale)
- 2 topics
- no comments
- no user posts
- related posts
- linkbacks
- no date sort
- no all topic (focuses on select blogs only)

Blogniscient
- 5 topics
- sort by date or popular
- blog directory
- no related posts
- no linkbacks
- no comments
- no user posts

Technorati Explore
- 1000’s of topics,
- no comments
- no user posts
- no related posts
- linkbacks
- no all topic
- sort by date or popular
- blog directory (sort by popular, date, A-Z)…Top 100

Corante Hubs (smaller scale)
- 3 topics
- no all topic (focuses on select blogs only)
- no popular sort
- no comments
- linkbacks
- related posts via Waypath
- no user posts
- tag cloud
- editorial blog

Findory - Blogs
- personalisation
- 15 Topics
- no date sort
- no user posts
- no comments
- no linkbacks
- no related posts

Topix - blogosphere
- no topics
- comments
- user stories
- related posts
- no linkbacks
- no date sort

Topix is also getting into the freshest top news from the blogosphere.

Seems to lack a choice to sort between date and popular…each story is tagged (by the system), but these tags take you to general news about this topic, not the buzz from the blogosphere around this topic…hmmm.
So it seems that blog posts are now included in the news topics…is there a way to look at a topic, and limit to just news or blogs or both.

The good thing is that you can leave comments to a blog post, and they are all aggregated into a Forum…there is also a User Stories section where you can submit your own stories, and others can leave comments…it would be good if you could click on the user’s name to see all stories from that user (kind of like inbuilt blogs…similar to Pluck Insite)…I don’t think these User Stories will make it to the front page.

Other news sites based only on user stories are Gather (content isn’t news focused), Commontimes, Newsvine, and others.

Another interesting thing is that under each post it lists related posts (they are just related in the way Memeorandum lists related posts, these are not neccessarily linkbacks)…if you click on sources (related posts) it launches to a page to show extracts from each source…including a cosmos type link would also further the discussion coverage.