From zero to five stars in Facebook, applying common sense Information science to a sluggish social networking giant
To begin this tale, let me start with your average search on a system like EbscoHOST for information you need about, say, carbon dating as it applies to erosion studies on the West Coast in America. The information science way, or one of them, is to allow you to formulate phrases and assign rules to how they fit together so that you can narrow that down: "carbon dating" AND (testing AND (erosion OR "soil loss")) AND America AND "West coast". The end result is that you get your data. Or, you get a negative result and you peel away layers. Lose the "West coast". Maybe change "soil loss" to include words like "washing away".
If you were on Facebook, you would have the equivalent of maybe specifying articles about carbon dating, or you might can get articles about soil erosion. Or, if you clicked down on the geographical relevance tab, you might find articles about the West coast. Combine those things into one? You've got to be kidding. Facebook does not care if you want to see the feed about a person AND only see certain aspects of that feed. So what if you want to find out what the hottie from your class said about Howard Zinn a week or two later. Click, man, and keep on clicking until you find it. This is because Facebook has made the number one dumb mistake ever in terms of information storage and retrieval: it thinks you can accomplish everything using single element delineations. You can search for stuff by a friend's name, OR by a type of application, OR by group of friends. It will all be in quasi-chronological order and access points will only occur in stream as opposed through discrete points.
The first solution, and really the best, is to allow pesky little ANDs and ORs to run around. If you have Gmail then open it up in your browser and do a search like this "from:WHOMEVER (monkey head)". That will find emails, assuming they exist, where whomever you typed in talked about monkeys and heads. Isn't that amazing. That's searching more than one element at the same time. That must be impossible black magic, Google. No way Facebook could do something like take data it already has (dates and app names, also basic word matching that it uses to generate advertisements) and allow people to use it at the same time you apply a user-name limiting field. That would be mucking about in stuff that no man should know.
Let's assume that multi-element filters are not going to happen, ok? At least not of that type I talk about above. Not as a step one, anyhow. Let us assume that Facebook would only use a system if, and only if, it was so simple that it's least common denominator could grasp it. Let's also assume that Facebook profits from a few extra clicks, and so any change needs to include a little bit of digging to get what you want. This is where I propose the five star method. Every friend is assigned a star rating, and the higher rating, the more about them shows up in your feed. Except, rather than make it some esoteric "Oh, I didn't think you would like that post" behind the scenes mumbo jumbo; let's start with the building blocks of Facebook itself: users and apps.
At five (5) stars you get to see everything about the friend. Whom they friend, what the became fans of, all of their application output, and the standard core applications of pics and notes and so forth. Then, at four (4) stars, you shed down to the application data and updates to their profile. You no longer see the book-keeping and mutually shared serendipity of whom they've friended and when. Three (3) stars, possibly the default, sheds that down to only the core applications: statuses, photos, links, notes. Two (2) stars gets rid of the notes and the links and only shows you things you would see at a glance. One (1) star is down to just the status updates. Then, at zero (0) stars, you see nothing in your feed. You have to go to their profile to see it. Make an option that says that 0-stars people will get bumped up to 1-star people after so long (with one option being never) just to bring them back in touch.
There you go, a simple two element system (application range and name) combined into a simple metric that could be grafted to the site with no more upset than their usual changes. Default everyone to 3-stars or 4-. Then let people edit. Where the "hide" link is now, change it to a 5 starred bar with a symbol for 0-stars.
I am half tempted to make a page on Facebook and see if I can get people to sign up for it. See if they could use this actual tool to make it easier, and more productive, to look at their site. At any rate, once they put that into place, a better multi-element method should be put into place for when I want to find out, say, when my sister and law has posted links to her blog. Facebook is eating up an increasing amount of our information generation budget. It is only a matter of time before our profiles take on increased elements of webpages and our Facebook mail supplants email. Without a better system in place, that will be a dark time where much of what we generate goes into advertisement algorithms and filters; but ceases to be useful for us. That's step two, though. Let's get step one done, first.
Si Vales, Valeo
file under ...on Linux & the Web
and (...on Science & Information)