Twine: A perspective on information
The Xpragmatic View #103
June 15, 2008
by Marc Buyens, Xpragma
Download as PDF-file 
The semantic web is all about getting better information in an easier way. Today, a first generation of solutions opens up for the public. However, we have still a long way to go.
As we mentioned in The start of the semantic web?, some two months ago we started using Twine. So it is time for a first evaluation.
To be clear on this, the current Twine product still is an invitation only beta release. The service is scheduled to open up for the general public somewhere the second half of this year.
In addition, Twine is a multi-faceted product and, as a result of this, also a rather complex product for the novice user.
Therefore, Radar Networks, the company behind Twine, still has quite a bit of work to do improving the overall usability and manageability of the solution.
However, that is not what we want to discuss here. Instead, we want to focus on our current perception of "the bigger picture" of this solution. To what extent will Twine be able to live up to the expectations? To what extent is this a step forward in the quest for the semantic web?
Bookmarks and tags
As for the Web 2.0 scene, we have always been a limited participant. We are member of a couple of social networks such as LinkedIn and Xing, but that's about it. We looked at the Facebooks, Diggs, Youtubes and Flickrs of this world, but they all seemed an enormous waste of time. We played a bit with Del.icio.us and Stumbledupon, but found no added value. So, now we are on Twine.
And we must admit, there is something that we like about Twine, but it is hard to describe. As we mentioned, we never really used social bookmarking solutions since they did not deliver any added value. We are not a professional collector of links or documents "in case that", so when we want to keep something, there must be a valid reason. We are also very systematic in arranging the stuff we collect and we do not change PC or laptop every two hours, so there is little need for a central repository.
Therefore, what we like about Twine is not that much the bookmarking and the tagging, but the fact that Twine invites us to think differently about information.
Categories, groups, folders and other twines
Within Twine, items (bookmarks, notes, documents, e-mails, images...) are "stored" in so-called twines. Although twines look a bit like folders, they are essentially a logical view on the available information.
When a user uploads an item to Twine, the item is added to his personal twine. The user can then create other twines (dedicated to a certain subject) and "share" some of the items with this new twine. This twine then becomes a view on all the items of that user related to that subject.
This concept then can further be extended since twines can be made public, allowing other users to access the content, to add additional items, to share items with yet other twines, to comment. This way, twines also act as discussion forums.
In some of the feedback given on Twine, some users have suggested extending this "structure" by adding things such as categories, folders and subfolders to twines. However, we do fear that they miss the point. This is not the way to go. Forget about folders and categories. Let there be information.
A perspective on information
Therefore, one of the biggest shortcomings of Twine is likely the existence of twines. While these might serve a "giving us a comfortable feeling" purpose (almost a category), they have little real utility, unless for closed groups or very focused discussions.
In theory, twines offer a perspective on information. Unfortunately, they are created by individuals, which means that they essentially represent these individuals' perception of the relevancy of the information.
As a result, when a certain subject is of interest to you, very likely there will be several twines that have something to do with this subject. Unfortunately, none of them will provide the perfect perspective. Your real interest will be a complex cross-section of a number of these twines. However, in order to capture this cross-section, you have to monitor them all.
Therefore, as someone suggested, the better approach would be to work the other way around. In such approach, you do not have to subscribe to anything, but Twine would learn about your interests by looking at the items you collect and the things you read. Building upon this understanding, Twine would then send you all the new items that emerge on Twine that match your interest profile.
Excellent idea, but with today's technology unlikely to be feasible, although Twine already makes recommendations about twines you might want to join or members you might want to connect to.
Therefore, a more realistic approach might be joining "searches" instead of twines.
Indeed, one of the power tools of Twine is the search capability. So, we are hoping to use this to our advantage. Today, we are gradually uploading and collecting items on Twine. Not many, since we only want to keep items that we have read and for some reason we want to keep.
We do not intend creating twines, unless for specific topics we want a discussion on. So, there will only be "stuff": a long list of uncategorised, but tagged, items. And then, we search. We are quite sure that we will find the information we are looking for. Simple. Great.
So, the same should be made possible for information discovery. Today, you essentially discover information on Twine by searching for twines related to certain subjects and then joining. This will give you a feed of new information, but as said, very likely too much of information that you have to wade through.
A better approach might be subscribing to search patterns. As described above, you find interesting twines by searching for certain tags, keywords, etc. So, why not keep it that way? Instead of joining the twines you found, you join the search pattern that you used to find these twines. Any new public item that is created somewhere that matches your search pattern will get into your interest feed. If your search is very broad, you will get a lot. If your search is very focused, you'll get the real stuff that (you think) is important to you.
Should not be too difficult...
Categories: Semantic web (Web 3.0)
RSS feed: http://www.xpragma.com/english/rss/xpven.xml
