# Flat tagging vs Hierarhical Ontologies -> what is in between? [closed]

I'm looking for different approaches about organizing information.

Could you provide reference to good overviews of approaches or do summary on your own ?

On the one hand there are flat tagging systems, on the other completely hierarchical ontologies. There are a lot of alternatives, variations and things in between.

-
 I don't understand how this question relates to computational science. Could you clarify what makes it on-topic for this site? – Geoff Oxberry♦ Aug 28 '12 at 6:07 @GeoffOxberry Sure. Organizing information is a kind of generation of additional meta-information, special kind of. Such meta-information might be defined and expressed in many ways, from flat tags (representing one namespace of sets) till graphs etc. I've tried to used simple language in question to make it more straightforward. So question is about such models of organizing information, representation of such models, dealing with collaboration between users. More specifically, any comparison, showing either adventages or disadvantages of given approaches, or mathematical foundations. – Grzegorz Wierzowiecki Aug 28 '12 at 9:41 Btw. it's a pity that I had no change to reedit this question before closing - I've been sleeping. I hope it might be reopened after improvement. – Grzegorz Wierzowiecki Aug 28 '12 at 9:42 You should still be able to edit the question in hopes of getting it re-opened. As it is currently posed, however, it doesn't fit well into the topics that are laid out in the faq for scicomp. – Godric Seer Aug 28 '12 at 12:11 In spare time I will try to figure out how to better fit. Thanks for responce. – Grzegorz Wierzowiecki Aug 28 '12 at 12:14

## closed as off topic by Geoff Oxberry♦Aug 28 '12 at 6:07

Questions on Computational Science Stack Exchange are expected to relate to computational science within the scope defined in the FAQ. Consider editing the question or leaving comments for improvement if you believe the question can be reworded to fit within the scope. Read more about closed questions here.

In Delicious I liked very much collections - a kind of special tags for tagging other tags. Thanks to that one tag could belong to many collections. This worked very well - great trade-off between mess of total hierarchy (ontology) and complete lack of it. One of discussions about collections on Zotero forum.

People working on Folksonomies discuss that hierarchies might be overrated and tag cloud might be enough or machine learning in order to perform some clustering.

What I like, are non conventional approaches, as I thing the solution should emerge from mixing many of current approaches in order to make something flexible.

Here is discussed concept of ordered tags, where flat tagging is extended with information kept in their order. On the one hand it give much more power to traditional tagging, on the other is not in conflict with standard concept at all.

Another innovative approach worth mentioning is Kleenk system for collaborative organizing research papers. It's uniqueness is organizing papers through relations between them, as it happens in our minds. Of course relations bring semantic information about meaning. We do not have collection of tags and tag things. We have collection of concepts (which might be things, tags, abstract, matter... whatever) and we connect those with relations. Kleenk.com is approach of providing way of doing so in visual collaborative way. I would love to see more tools like this for other stuff (we sites, music... everything), maybe combined with previously mentioned approaches.

As summary I might try to list those in list from flat to hierarchical structure:

• Flat Tagging
• Folksonomies
• Ordered Tagging (Flat tagging extended by ordering)
• Collections of Tags
• Tags with Inheritance
• Hierarchical Tagging
• Tags with multiple inheritance
• Graph of Tags - Relations
• Graph of Tags - with relations consisting meta-information (like semantic)
• Ontologies

Where do you put which - it's up to you. I believe there must be works on comparing those or showing some properties (maybe formal homomorphisms, isomorphisms). Those are just systems of organizing information.

-