So for Project Greenshift, we talked last time about introducing non-linear ways of exploring content that are controlled and influenced by the individual user rather than a strict algorithm. We achieved this by giving the user different options of content to choose, but now we need to select different relationships between content to give them the options in the first place.
In selecting these relationships, they should be local to the current piece of content. This means that we should be able to easily traverse the blockchain along this relationship to get the piece of content and this traversal is dependent on our starting point. Depending on your starting point, you get different results. This approach ideally gives us a little more diversity of content than picking topics or organizing by the nebulous "trending" or "hot" algorithms.
One relationship could be the newest piece of content by the same author that has yet to be visited. You simply search the author's posts on the blockchain and pick the latest one. A simple relationship but one that could yield interesting content if you happen to enjoy the author you are currently reading. That being said there are other options, so if you don't like the author other relationships exist. And that's where some of the power of the approach lies.
Another relationship could be the most recent piece of content that the current author has upvoted. This allows you one to traverse between authors without needing that author to resteem everything they like. Granted the content an author upvotes may not be related to the content they produce, if there is overlap that allows users access to authors in a particular area or style that might not be discoverable through the conventional tag system. Also if a particular author is selective about how they upvote, you are bound to get better results by using their upvote patterns as a reference.
Piggybacking off the last relationship, we can use the most recent resteem in the place of the most recent upvote. In this case, the content is more likely to be related to an author's area of interest since they are introducing to their feeds for all of their followers to see. But perhaps there is less potential for exposure using this option.
As you are bound to notice all of these relationships use the most recent article in terms of the general relationship. The reason for this is the 7-day payout window. Obviously there will be content worthy of interaction outside that window, but initially the tool will be the most useful to get smaller authors more attention and give audiences content that they can at least have some chance of getting curation rewards. Granted I'm not a big fan of chasing of ROI, I still feel focusing on content within the seven day window will be important at least initially.
One last relationship to consider is a little more complicated, but one I've just thought up right now. Consider the people an author has upvoted in the past seven days and find the piece of content for which the author's upvote contributes most to the overall weight of the post. A complex relationship, but perhaps a meaningful one. What does such a relationship reveal to us. It means that the author has found a relatively undiscovered piece of content that they like a lot, or they are the sole givers of an upvote to that content.
The potential of this relationship to uncover unknown and obscure authors via distributed trust in the current author is actually a particular interesting area I would be excited to uncover in more detail. And that's just the tip of the iceberg in terms of these different relationships we can define on the blockchain. Quite exciting!
But before we select the relationships that are to be used in the early stages of Project Greenshift, there is another thing that we must consider first. A starting point. Now that we have a way to traverse the blockchain and have begun to consider relationships that lead us to potentially interesting and novel content, we need to have some initial seed to begin the whole process. And that's what we'll consider next time.
This post has been just added as new item to timeline of Q-Filter on Steem Projects.
If you want to be notified about new updates from this project, register on Steem Projects and add Q-Filter to your favorite projects.