Weighted, parented tag-tree

A unification of the tree model with the bucket model for searching posts

Bucket model means: slap a couple of tags on a blog-post and you can search for it. The post is essentially thrown into a bucket, where you can go fish with tags.

Tree model means: the default folder structure for every computer.

I now propose something in the middle, that would work better with the results of user-generated content.

For any post, we force the user to put it somewhere in the ever expanding tree. A user can also create a new folder.
A tag-weight on a post, will consist of a number of points and a total of points, that has been awarded in relation to this tag/post.
a couple of things will happen when he or she does that:

• The post will be created, it needs to have a unique name, in the folder.
• A tag will be created (if it doesn't exist) in  the tag cloud.
• The tag will be a child of another tag, all the way back to a root-tag.
• The tag will be added with weight one to the post and every parent tag will be added to the post, but the weight of this tag parent tags decreases linearly to 0 for the root.
• Essentially it is given points for the depth of the tree of each tag, the total given points is collected and the weight is those points divided by the total.

This gives us an initial weighted structure. Where weights are stored as points and totals.

Tree management:

• Every user loads the tree only once.
• New additions to the tree will appear to the user, as posts with a suggested placing (tagging)
• The user can either discard the post OR accept the proposed tagging (as most will) OR move the post to a different place. (this will reinforce the weighing by 0.1)
• This place will be permanent for this users copy, but will also result in:

1. The weights of tags on the post will be amended  with the new tag by this user.
2. The weights will be amended all the way back to the root, which will have a weight of 0 or not be a tag..
3. The new weights will be the points divided by the new totals.
4. The popularity of the post will be measured by the amount of readoptation. (Also a point/total system)

This will give us a system in which every tag is either searchable by term, by tree AND freely moveable for each user. Each user searches only in his own copy of the tree, but the NEW copies are subsequently more weighed and a consensus product.

Also it allows for more intelligent tag searching, because every tag is weighed.

If users loose their copy of the tree, they can always get a copy of the newest tree and prune and modify it.

By using a competition system in the posts and artificially limiting the amount of posts to the top x adpopted ones, the tree prunes itself.

Any deletions from the main tree may also be suggested to the user and the user may then follow by pruning his/her tree (deleting the post). (this will have a smaller impact on liking)

A user can also at any time prune his/her tree and this will have a bigger impact. (it is essentially a full dislike or a point less in the readoptation total of this post.

This technique can be combined with the previously proposed technique (2016) to evolve melodies. It can also be used with the proposed technique to have user-generated sounds compete for space in an evolutionary system. (Beep-o-tron-file-system)

And I totally forgot these posts were supposed to be in Dutch.. Maybe I'll translate it someday, not now..