Neurorganon - Personalized Topic Map based on bookmarks

[athanassios]

Hi Wandorers,
first please let me congratulate the development team for such a great piece of software Wandora is and for making it free and open.

Recently, I have been researching again on social bookmarking (tagging). As you probably know most of the software development efforts are centered on hierarchical structures with folders containing bookmarks (i.e. Browsers and Bookmark managers) and tagging techniques most often met at social bookmarking sites (i.e. delicius, diigo, faviki, perltrees, etc).

I have realized that Topic Maps and semantic technology in general is bridging the gap for these two techniques above. I do have a personal collection of about 6000 bookmarks classified hierarchically into 900 folders (i.e. topics). I have started experiencing serious limitations on navigating and updating this tree !

Therefore I started working on the "Neurorganon" bookmark project. There are a lot of workpackages I can think in such a project and many resources to utilize in a complete system. I believe Wandora is one of the best tools available for free to commit to integrate and develop.

In fact I am aware of a such a completed system based on topic maps, it is called fuzzzy( Leipzig University - Institute of Informatics project). But it is frozen at the moment.

Are you aware of such similar efforts, why the software industry or research academia is not interested in such a project ? Is it only simple tagging social networking that has captured the interest of the users and communities ? Perhaps I should search more on the subject, but if you happen to know more on the topic and you would like to share, please do so....

[akivela]

Hi Athanassios

Thank your for sharing your idea. Do you have any public interface to your Neurorganon project?

Thinking of tagging systems scene, I have a feeling that it is divided into two separate camps at the moment. There is a tribe of controlled vocabularies that uses well (or not that well) structured vocabularies for tagging. And then there is a tribe of folksonomies that delegates the work of structuring the vocabulary to their audience -- or just don't care the hierarchies in the vocabulary. I share your observation of controlled vocabularies. They really use tree-type (or folder-type as you write) of relations to arrange the tags in the vocabulary.

Kind Regards,
Aki / Wandora Team

ps. I have an impression that fuzzzy is a project of Roy Lachica. I think he is not related to Leipzig University nor the Institute of Informatics anyhow. Maybe someone knows better than me.

[athanassios]

Hi Aki,
thanks for the reply, indeed Fuzzzy is a project of Roy Lachica. I read his presentations recently. Fuzzzy is also mentioned in a Tutorial Introduction to Topic Maps by Lutz Maicher, Leipzig University and that is how I was misguided . There are a lot of unique features in that project about tagging and the use of topic maps. Unfortunately I cannot test it fully to see all user interactions as they have suspended sign up. If there are exceptions on the rule I will be glad to comment or participate on that project.

The key point I believe for the future of bookmarks-tagging systems is to allow any user to maintain his/her own bookmark Topic Map by merging or use part of standardized/authoritative Topic Maps on a certain acknowledged subject e.g. semantic web. Certainly using controlled vocabularies for tagging is a step ahead towards that direction but Topic Map technology is far more rich and appropriate to model such as task. Therefore I believe that developers or social bookmarking sites have to give the users the ability to i) maintain their own personal bookmark (Topic Map) structure, ii) share and collaborate based on the Topic Map technology.

I am given the impression that the broad scientific community, software developers and others have not fully exploited the power of Topic Maps and the effect it can have on social networking. Unfortunately I do not have any resources to build a commercial system, and I am not employed by anyone at the moment. But I am still very interested in the conversion process of building a personalized Topic Map from my personal bookmarks collection. I will try to make public the result in the neurorganon site.

I am at phase 0, where I am looking on ways to standardize, i.e. create automatically or semi-automatically subject identifiers (names) and occurences for my topics (bookmarks folders) based on semantic vocabularies (umbel, opencyc, freebase, dbpedia, yago2, wordnet) I think these are the most popular at the moment !!!

OK if there are others interested in contributing we can make it a nice open collaborative project. Regarding to the user interface, graphics part I strongly recommend adobe flex. There are a lot of powerful and open software tools for development there that I am aware of.

jacta alea est

Athanassios

[akivela]

Hi Athanassios

If you feel that Fuzzzy would do the trick, I strongly encourage you to contact Roy Lachica. I don't know him personally but remember reading Fuzzzy was invaded by spammers and he had to limit the user registration. If you ask the user account personally and explain your motivation, I assume Roy just can't refuse creating an user account to you. But, as said above, I don't know him nor his plans for the fuzzzy. But it is definitely worth of a try.

If you use Wandora and feel it misses some important and general features you could really use, please, do inform me. For example, I recall that Wandora really hasn't import/extract feature to transform any bookmark-file-formats to topic maps. Also other Wandora users might be interested in such a feature.

Kind Regards,
Aki / Wandora Team

[athanassios]

[quote="akivela"]If you feel that Fuzzzy would do the trick, I strongly encourage you to contact Roy Lachica....[/quote]

Yes, I invited him to read the messages we exchanged and to participate in our discussion.

[quote="akivela"] I recall that Wandora really hasn't import/extract feature to transform any bookmark-file-formats to topic maps.[/quote]

I think in principle any feature that creates, reads, updates, or deletes data in XTM should be in the form of a web service. Moreover it would be even better if those services can be modularized and interconnected in a pipeline fashion.

For example, a web service that takes a bookmark from the web browser and converts it to a topic, another one that assigns identification, another one that asserts specific statements (names, occurences, associations), another one dealing with scope.... etc

Perhaps a scripting language javascript, python, etc would be ideal to glue these services together in conjunction to TMQL.

Another important point is that the user has to be in control of his personalized topic maps and exchange information through his own perspective on these topic maps he created.

Other topic maps should be standardized or be more authoritative by communities of experts and modifications there should follow strict rules. Perhaps this way the redundancy and spamming of information will be reduced significantly as these actions will be confined.

[roylac]

Hello Athanassios
Please read the blog post I published today. I believe you will find some relevant information there.

http://fuzzzyblog.blogspot.com/2011/09/ ... ocial.html

I'm currently deciding if to discontinue fuzzzy or if I should re-architect it as it has not gained much attention or has proven to be a very useful system. That is why you can't sign up. If you would like to try it out let me know and I'll create a an account for you.

If you create a wish list or specification for a new bookmarking system I might develop that into a new version of fuzzzy. Anyway a new version of fuzzzy will not have a TMQL interface as I am now more focusing on RDF technology rather than Topic Maps technology.

Best regards
Roy Lachica
http://koios.org