How to solves the flatness of the list of tags?

Currently, all tags are at the same level. The growing list of tags makes it difficult to browse efficiently.

A hierarchy of tags is provided independently as a set of trees. It can be build on the fly (incoming new tags are put somewhere in the tree organization as they come) by “demexp organizers” or existing large scale trees can be used (for instance wikipedia has some form of global tree organization).

This is only an organization of the tags, not the the questions. Note that that a question has usually several tags so a question can be considered to belong to several nodes/leaves of the tree at the same time.

In the left field of the demexp client (”Tags” field), the tags would be displayed as a list of organized hierarchical trees. A node or leaf that would “contain” questions that are N (New) or U (Update) would have a specific color or font. When opening the demexp client, the trees would be opened to show the nodes/leaves that are N or U only.

It would be too much trouble as a first stage to extract a relevant tree organization from independent source such as wikipedia, so a software should be written to allow tags organization. Anyone could propose its own tree organization of tags, but demexp can provided a default one, built by people elected within demexp.

None

When someone launches the client, the client:

• reads the hierarchy of tags and
• displays it in the “tags field” in the form of a list of trees with fold/unfold buttons for the nodes. A tag that has not been organized yet can still be displayed as a tree with a single leaf.

The client “preference” menu should allow you to specify an adresse for the storage of the “organization file” so that anyone can propose one.

(Please note that when delegation will be implemented, this will have to change slightly since delegation and hierarchy of tags will interact: if someone has delegated on different delegates on 2 tags assigned to the same question, he/she will be asked to set a hierachy between these tags. This may have an effect on the tags organization. As a consequence, tags organization will become a mix of “existing organization” and “personnal organization”. Therefore, when the client downloads the organization file, it will have to be merged with the personal one, and personal organization should previal when there is a conflict.)

This software should:

• read the existing “organization file” where it is located (typically stored in on the server)
• read the “flat” list of all tags and identify the new ones (just compare between flat list and organized tress)
• have a GUI (preferably) to allow the user to modify the tags organization
• write the new “organization file” where it should be located

It does need to be have fancy GUI because it is targetted at a limited audience.

Examples of used classification:

1. Dmoz : http://www.dmoz.org/
2. Wikipedia