I am well aware that everyone on this site has invested some time and energy, and I thought it might be appreciated if I gave a bit of insight into my thinking on sites like Hubski. I have both philosophical and technical motivations for what I am doing; however, the following focuses primarily on the technical ones.
News aggregators have a challenge, and I don’t think it is being approached it the right way. There are many ways to define this challenge, but it can be summed up as this:
The content I want to see should be readily available to me, and the content you want to see should be readily available to you. This should hold true even if you and I prefer different types of content.
For aggregators that have a limited scope of content, focusing upon content quality provides a pretty good experience. Pre-defined expectations for the type of content exist, and user voting focuses more upon quality of content rather than upon its type. For example, most users on Hacker News would not upvote a posted song by Prince, even if they had an interest in the artist.
On the other hand, on general news aggregators, both the quality of content and the type of content are open questions left for admin and users to answer. As these sites become popular, these two questions create a tension: although users might agree that a given post is of quality, they might not agree that it is the type of content that they most wish to see. Over time, this problem has been addressed in different ways. Here are some solutions that have been tried:
1) Give some users more influence than others: Users can be given moderation powers, their posts may have more staying power, or their votes might be given more influence. The benefit of this type of moderation is that it reduces the burden on site admin to curate content, and it rewards heavy users for their contribution to the site. The downside of this type of moderation is that it can make other users feel like they have less influence upon the site, and if these powerful users have significantly different opinions on content or quality than a particular user, the experience for that user is diminished.
2) Break up the content: The aggregator is divided up into multiple sub-aggregators. The benefit of this approach is that, like a content-specific aggregator, the question of which content is desirable is more-or-less settled. One downside of this division is that it divides the audience as well. If a subdivision of the aggregator is obscure, then few will see it, and the chance it will garner discussion is reduced. Another downside of dividing content is that divisions can be redundant. For example, engineering might overlap with technology, which might overlap with science. As a result, the decision about where to submit or where to find content can be somewhat arbitrary.
3) Active moderation: This type of quality control is usually achieved by moderators or algorithms curating content. Certain sources might be banned, some content deleted, or some users might be restricted or banned. The benefit of active moderation is that a consistent vision or tone can be maintained. However, as bias in active moderation can narrow content, this works best when applied to an aggregator of limited scope, or applied to a sub-aggregator. The downside of this quality control is that it is only as good as its moderators. If moderation is not predictable, or seen as unfair, users can feel restricted, or in the case of content bias, serendipity can be reduced.
4) Popular consensus: This type of quality control is usually achieved by a voting mechanism; users vote a story up if they like it, and in some cases, can vote it down if they don’t like it. The benefit of popular consensus is that the content that is seen is that which most users prefer. However, the downside of this type of quality control is that users that don’t share the majority opinion have a diminished experience.
Ideally, a user could see all content submitted to an aggregator. However, this is not possible. To solve this problem, aggregators typically try to focus the content, and improve the quality of the content that any given user does see. Current approaches are good, but not great. There is vast room for improvement in the aggregator space, and there are real problems to be solved.
Aggregators provide value not found in other sources. One value that aggregators provide is community, another is crowd-sourced information analysis. Although aggregations have been around for some time, I believe this is only the beginning.