In the body of our proposal we have limited ourselves to a select few distinguishing features because:
Any endeavor must have a limited scope in order to be achievable at all, even greater limitations must be implemented to ensure it is achieved in a timely manner.
We wanted to keep the body of our proposal to an easily digestible serving and including the kitchen sink would not have furthered this purpose.
That said, we have a number of other "ideas" sitting in the wings which we hope to be able to implement once the base service has been established and is growing successfully. We've outline a few of those here.
We believe that an analysis of page content can help us determine the similarity of pages in the information they present. Rather than showing every result with similar content at the top of Phoebe's results we can make each result expandable with the very similar content listed beneath.
This will allow searchers to access all of the content available on a topic but will also emphasize content which provides unique qualities. Usually one does not need near identical instructions on a subject from multiple sources. Instead one may desire a brief introductory article, a more extended treatment, and perhaps several alternative views.
Often these categories get buried under pages of high-ish quality but duplicative content. Using this relational analysis method would allow us to surface more types of content to the user.
We believe that understanding of the other improves society. Today many search engines capture users in bubbles in which they are primarily shown results that reinforce their own presuppositions almost exclusively.
We believe that by using sentiment analysis we can provide results which are not only sorted by quality but also include diversity of perspective.
We are unsure of the exact implementation of this functionality at this time though it may look like a tabbed interface for the search results which allowed one to switch between the best results for various sentiment positions.
The number of spelling, grammar, and factual errors present on websites, even the best of sites is appalling. We believe we have a solution that will allow website owners to receive and (almost) automatically apply valuable updates to their content from their audience.
This would involve a browser extension on the end user's side which would allow for client side editing of third party site's (at least text). On the server side the CMS would utilize a plugin which would allow website owner's to review recommendations and if found to be valuable automatically apply the updates to their content.
This system would be another use for The Trust Network and another way in which users could gain authority on Phoebe.
We believe this could also open the way for paid crowdsourcing of materials that would significant improve content quality and clarity.
The Internet Archive has done humanity a vast service with its Wayback Machine which archives much of the now dead web. As avid fans of Wayback Machine we'd like to see it used more widely and believe that we could find ways to improve the discoverability of materials and perhaps assist in making recovering these materials more speedy.