It's about time: leave your box, follow the links

How many web accounts do you own? Do you use a public email service like GMail or Yahoo? Do you have a LinkedIn profile? Did you ever notice that you are dealing with the same projects, people, and topics in all these applications?

I stopped counting my online accounts at 44. I guess that I now use over 70 systems on the web. Google for Leobard Sauermann to find some of them. I lost the overview. My accounts too - they lost track of each other, like a large family. My Flickr account does still talk to its brother - my private Blog - from time to time, but it does not know of its far cousin, the LinkedIn profile. Some of them integrate with their godfather Google, but Google does not integrate with all of my 70 accounts. For me, this is a problem. The high-strung reader may see a simple solution:

"the problem is the many accounts - use only one software".

Not possible. I love all these systems. Another interpretation by an excited web developer may be:

"the problem is service integration - make a mashup for the 70 services".

This is also not trivial, because you need to connect each of the 70 systems with all the others (my school math gives me 2485 pieces of glue code to interconnect 70 systems). If you come up with another solution to the overall "many boxes" problem, add a comment. Fixing it the excited, high-strung way only treats the symptoms. But for the clan of applications I use - I am looking for a cure, not only for a medication to dampen their symptoms. So, what is really happening here in the world, and what is the solution? I propose a way out:

  • Keep using your different software systems, put your data in these boxes.
  • Let our semantic personal assistant software Cluug help you connect these boxes with links.
  • See how everything sticks together, get an overview

Desktop and mainframe applications used to store data in hierarchical file systems. The world was made out of boxes: each box contains more boxes or more files. Computer systems were fine. Each system had its boxes. Then the world wide web came and applications and data were everywhere. A precious few people started to build hyperlink-based web software [such as Xanadu] in 1965 and were stunned by the possibilities and promise of hyperlinked information systems. These pioneers saw a glimpse of what is to come and tried to build it. On hindsight we now know that it took 40 years and a whole industry to realize their dream of the web. Still the dream was the same: we had boxes, and hyperlinks were the innovation that opened boxes. They allowed us to jump from one system to another, in a web.

The web is here now, and ... Yes, We Can Change ... and walk the way of the web. One example shows how we jumped out of boxes and followed the links:

  • The Google search engine understands hyperlinks. One way of knowing that a page is important for a topic is the number of hyperlinks pointing to it. Hyperlinks bring structure into the chaotic web.
  • The hierarchical "boxes and trees" are not important anymore.

The web is an uncontrollable big linked vibrant place, with endless interlinked islands of knowledge in many many systems. Trying to categorize or order these boxes is a fruitless task. If you study the history of Google [nice infographic], there was a time when search engines tried to put everything into boxes: the Yahoo Directory, and the effort Google did when they bought the DMoz directory and featured it on their front page. It was removed later [because boxes are not important anymore]. Google uses many ways to give you search results, but one major contributor is the well-known "page rank" algorithm using hyperlinks as data basis. Google understood that links are as important as hierarchical directories.

At Gnowsis, we do not address the whole web. We are helping you personally to get organized. We understand what a personal web of information means for you. For many of you, daily office work consists of many systems and boxes where you keep your data to get your work done.  You are free to chose whatever system you want, every box is good and purposeful. What we can do for you now, is help you make links between these systems and boxes. We know very well, that current software "boxes" don't give you links, and many examples prove that these links between your own personal data systems will be very useful for finding information. Our product Cluug allows you to link from a webpage to a project, from a project to a task, from a task to a document you are writing, and from the document to the project it is about. Your mind understandsthese associations in both ways: when the document links to the project, the project should also link back to the document. Cluug does that, it associates in both ways, like associative thinking.

Google, Microsoft, Yahoo, and many other future-oriented companies are members of the W3C consortium to agree on standards that make the systems work together. Since 2004 (when our company Gnowsis was still part of DFKI and the University of Vienna and we focused on scientific research), DFKI and DERI and other W3C member institutions started to forge new standards for personal information management, now know as the NEPOMUK standards. We know that you need something to link between your boxes. We know that such systems take decades to build, so we are building on the shoulders of giants, namely the Semantic Web and NEPOMUK standards. What you get now with Cluug is a way to jump from one box to another, following the links. In the coming years, this will allow you to do many more things. Because I trust standards more than proprietary boxes, you will be able to keep your links for a long long time.

Still, there is nothing new under the sun. The idea of a semantic personal information system keeps coming up again and again - in 1945 Vannevar Bush called it "the Memex". If you have ideas what you would like to do with your links, comment here, new and old ideas welcome! And please promise me: if someone offers you a system - check if it can follow the links.


btw, this article was inspired partly by a discussion we have over at the German "Enterprise 2 punkt 0 Blog" about Semantics: