Actually Raphael size has nothing to do with why Facebook search is so bad. Google has indexed 30 trillion pages and it still returns results that are accurate and relevant. Facebook search is and has always been bad because it cannot deliver relevance, it has only just go to grips with disambiguation and it still knows nothing about accuracy. In order to deliver accurate, relevant results you do not need to just index posts, that’s just data. You need to then connect each point of data and understand its importance in relation to every other point. That is actual knowledge which brings us to Google’s semantic search and what it does. So an algorithm does not just weigh what a post is in terms of your search query — that’s Boolean programming that works on statistical probabilities based upon search term frequency (and placement) within a particular post, it weighs the meaning of the search query in relation to the searcher’s profile and past search history in order to understand intent. Facebook, historically, has not given a hoot about searchers (or members) and their intent and it utilizes its Facebook’s Graph to limit the data set from which the search query can draw information on the principle that you are not really, within Facebook, searching for knowledge (you go to Google for that) you are searching for whatever your friends have shared that might be relevant to you. Facebook’s search is also driven by a semantic search algorithm but it looks at the connection between people as opposed to the connections between things.
The difference is an important one. This means that unless a member is active within the wider set of connections within Facebook then their search experience is likely to be poor. Facebook search was intended to help people find and add more friends utilizing the friend-of-a-friend connection and the reason it is focused this way is because what is of value to Facebook are the relationships between people (which can be used to advertise and recommend products and services) and not the ability to discover knowledge or find something truly useful, independently.
This is also the reason why, while every search engine works to introduce serendipity in its programming, intended to broaden the search window for the searcher, Facebook introduces a filter bubble which forever narrows it. This narrowing which is anathema for anyone living on the web is, from a Facebook perspective, a honing down of likes and dislikes and a clearer understanding of interests. Its purpose is not to join a billion people globally and let them interact freely, gaining knowledge, experience and perspective from the connection — for that you really need to go to Google+, Facebook wants to corral the world, putting people in little boxes and pumping endless ‘targeted’ ads at them based upon its explicit understanding of their interests, drawn from their tiny search filter bubble.
Good search is hard to do well from a programming point of view. But it is also hard to do well from an end-user perspective. One must truly want to liberate information and democratize access to knowledge. Facebook has never been about that nor is it where it’s at right now.
