Showing posts with label liquid browser. Show all posts
Showing posts with label liquid browser. Show all posts

Tuesday, March 20, 2007

Liquid Browsing - a new paradigm for accessing huge collections of data ?!


As promissed yesterday, I've installed iverse's liquidfile, a kind of substitute (or supplement) for the mac finder. (For all those mac illiterates: the finder is some kind of file system browser). Liquidfile applies the principle of liquid browsing to the filesystem of your computer. To get a short overview, how liquid browsing works, just take a look at the following short movie presentations [1] [2]. It's rather difficult to describe the way how it works with words, because its a visual way of browsing...and thus it is better explained in a visual way also. It's a 2 dimensional visualization, where e.g. the x-axis represents a timeline (as e.g. file creation date) and the y-axis represents all files in an alphabetical order. Then, every file is put on that grid and denoted with a bubble. The size of the bubble (which itself is semi-transparent) reflects the size of the file. The nice thing in general about liquid browsing is that you are able to visualize huge amount of data also on small displays. In that case the mouse pointer acts as some kind of magnifying glass.

As for the file system on your computer, this kind of visualizations has some advantages. First, you can dissolve the entire directory structure (if you want) and look at all files at once (on my computer this were more than 15.000 files). You have the possibility to combine complex (realtime) filtering with several selection mechanisms to find the files you are looking for. Simultaneously, the order in which the files are presented in general stays always the same and does not change. Thus, your visual memory will always be able to memorize the approximate position of a distinct file that you are looking for.
Of course the system has several drawbacks (at least now). My macBook is rather new (january 2007). Thus, almost all of my files have the same creation date (most of them were simply copied to the macBook on the very same day....). Therefore, the time axis will work best, if you start to work with your computer and they will be created/changed over the time.
Another drawback is the limitation of the axes to represent only filenames, filesizes, or creation/change dates. I would like to order files also according to other (also content based) criterias to get a better overview.

But, it's a nice appetizer anyway. What I would like to try out is to visualize relationships between entities (files/documents) with this technique. Just imagine bubbles of different entities and if I focus on one, all other bubbles that represent entities, which are in a certain relationship to the entity being in focus will highlight or move closer (of course in realtime...).

Monday, March 19, 2007

CeBIT Day 3 ... finally


Finally I'm back home at the HPI in Potsdam. Yesterday (Sunday) was my last day at this year's CeBIT computer fair and there were a lot of interesting things going on...
First, by chance I met HBS-fellow Jan Schmidt. Together with Oliver Gassner and Reinhard Karger as host he was talking about the use of networking platforms such as second life, their use, the hype, and social spam (read more about it in Jan's blog).
Another very interesting talk was scheduled afterwards: Carsten Waldeck, founder of iverse was talking about liquidfile - a file browsing application based on liquid browsing. Liquid browsing offers the possibility to arrange and manage a huge number of data items (e.g. documents) in a rather efficient way by a smart 2d-visualization. This principle is now applied to file system management on your local computer. There is an implementation for the mac and I will test it asap to give you a report.
In addition Carsten Waldek and host Reinhard Karger were talking about the plan of installing a social networking platform for archiving intellectual content (not in the sense of intelectual property). But, you should be able to express an idea, to upload your idea in any form of document, and to receive an official timestamp. Of course maintaining a database of 'ideas' will be rather difficult. Thus, all those ideas have to be translated into 'logic' (means with the help of ontology and rule representation languages). Then, if you have a new idea, you can check, if anyone else has already thought about it and if there are some people with similar interest. The other way around, you will be able justify that you have a claim in some idea because you have registered it previously. The trouble is....translate a natural language document into logic representation...you know that this - if done completely automatically - will fail (at least too often....). Thus, we will have to wait, until the semantic web will come.
On the other hand....we have been thinking and thinking again about possible killer applications for the semantic web. This(!) could be one....

BTW, I also met another HBS-fellow Andreas Scheper after the talk (read more about the CeBIT future talks and see some videos in Andreas sein Blog). We decided to meet again at re:publica in berlin, where als the HBS has a so called plugin that will be filled with some fruitful discussion...

So, CeBIT is over for me now... If you want to stay informed about ongoing CeBIT work and CeBIT after hours, keep on reading Miss Marple's Blog.