Musings on education, techology, and life..
“From what I knew of learning to write I asked Harold Bauer if it wouldn’t be possible to learn to play by playing tunes from the beginning without preliminary finger exercises. He cheered me with the assurance it would. Many second raters present were scandalized. Children are learning now without finger exercises. Think how much easier their education is to listen to.”
–Robert Frost
from Notebook 26
In conversations with George Siemens yesterday about personal networks, the question of network size, accuracy, and trust came up… in particular the concept of “friends” in existing social network applications. The issues aren’t new. To list just a few fundamental problems:
I’ve long preached that one of the most significant future values of creating and maintaining a social network comes from applications still to come which will makes use of one’s existing network to create new and more powerful mechanisms for finding, handling, and assessing information. The obvious first step, which is beginning to appear, is using your network connections (people and resources) to order and/or filter search as well as provide semi-intelligent recommendations for material related to those searches.
a) For further, more sophisticated use of our network, we need that network to have a more sophisticated understanding of relationship types (including negative ones)! In real life I have friends, peers, institutional colleagues, soul mates, acquaintances, bosses, immediate and extended family, etc. I have connections to people that I know and do not know, that are generally interesting and that are interesting only in particular contexts. There are people and information resources that I expressly don’t want to hear from and whose influence should have a negative effect on reputation, ranking, and recommendations. Granted, that no scheme will be perfect and leaving the creation of the system completely up to end-users might be too unwieldy… but it’s hard for me to see how an imperfect change could be harmful.
b) Beyond generating pretty graphs and impressive-looking numbers, the friend-of-a-friend network– even with a binary conception of friends– could already do much more than it does. I don’t need to know that through my connections in System A I am linked to a city-sized population of unknown people also in System A… nor is manually browsing that list likely to be fruitful. How about using some mathematical mojo to make use of the actions, information, and connections of that population to bring new information, resources, and people to my attention. What I see happening now is rudimentary at best, non-existent at worst.
c) If there is some scaling conception of trust in connections that can be expressed by value, at least, and preferably also by type or content, the door is opened for leveraging one’s network connection to take into account explicit, inherited, derived and even predicted trust. Operationally, trust is almost another social type… it might even work, at least as a start, to conceive of it that way.
d) Twitter makes some of the consequence of the single friends model explicit… and calls into question the model of almost complete transparency in a social system. I know I’m not alone in imperfectly and inconsistently fighting a reflexive feeling of obligation to follow those who follow me, which just isn’t workable for me even at my current levels of following about 30% of those that follow me. Making the model transparent so that we all know who all followed and followers are provides avenues for introduction and discovery… but it also makes the act of following into an uncomfortable social signaller confusing the act of instantiating a network connection– which can come in many types– with friendship in the traditional, social sense.
I don’t see a downside in enhancing connection mechanisms for those who choose not to engage them– those who want to simply have people who are connected or not with no concept of type or trust can do so– the calculations to be made would simply rank all connections equally as they do now.
Is anyone working on these issues? Is progress being made?
I had a dream last night that the US Postal Service announced it would start adding a $2 charge, payable on delivery, for receiving NetFlix DVDs in the mail. The announcement, written in microscopic print on the back of my bill cited the high volume of NetFlix mailings interfering with "normal" mail flow and delivery as the reason for the charge. Other services would not, at that point, be affected.
If I had a problem with this, the mailman (moonlighting from his job as "Speedy Delivery" guy in Mr. Rogers) pleasantly informed me, I could choose to have my DVDs tagged to eliminate the charge but it could take 2-3 weeks for me to receive them. Or, he said with a twinkle and a tip of his cap, I could write to NetFlix to see if they wanted to strike some kind of deal to speed things up a bit. I also noted that I could switch to a different DVD mailing service and receive the same movies in the same kind and size of envelope. At least for a while.
So goes net neutrality, without which the carriers discriminate, double-bill, and surcharge based on the type of content that is being carried, punish successful enterprises, and create a lever by which they can enrich themselves through sweetheart deals and collusion.
Net neutrality doesn’t interfere with tiered pricing by speed or volume or adjusting based on the cost of doing business geographically. In my dream, neutrality wouldn’t prevent the USPS from creating a new pricing for all mail of a certain weight or for allowing people to choose to pay for a special delivery method… as long as it didn’t discriminate by content or brand. The post office can’t refuse to send or delay my letter because it espouses Buddhist rather than Christian principles or because it is going to the Church of Scientology instead of the Church of God. Why would we give Internet access providers that kind of power?
