Signal vs Noise

Picture taken from Thomas Vander Wal's presentation, Granular Social Networks.

Thomas Vander Wal recently posted a great short presentation, Granular Social Networks. In it he tracks the complexity of relationships within and across networks, making interesting and important points about the overlapping of interests and following behaviours between connections. The only thing that I'd be keen to stress a little more would be the relatively haphazard relationship most followers necessarily have within social networking service relationships. While most of us have very few connections that we engage with across the entire range of their interests, activity and expertise (stalkers, the love lorn and private detectives aside). Similarly, even amongst those connections that we have an identifiable interest in - for example, I'm interested in your music consumption and recommendations - it's not usual to keep track of every single recommendation or playlist. There are just too many other things going on. So to a certain extent I don't believe that greater control - i.e. finer granularity within network channels - is the answer ( & you can check out my post on in service granularity here for further elaboration). While intellegent and sensitive service design, along with user digital literacy are important, a philosophical acceptance of serendipity and a kindly understanding of the human limitations for data absorption are also useful.  In the words of a Jaiku conversation I had with Terry Madley earlier today: "or maybe, only learn not to mind so much about the inevitable periods of disconnection. It's kind of good to not think about the info streams as if they were linear, let alone might constitute any kind of linear narrative. Maybe this is one of the reasons why lifestreaming is popular - the illusion that if you could somehow keep track of everything, there'd be a coherent story at the end of the rss rainbow."

The other issue that Thomas touches on is another of my current bug bears - signal v noise. I wanted to post here primarily to put on the (blog) record that both signal and noise are entirely subjective concepts. They aren't even stable.  What's noise to me on Tuesday morning might be be signal from heaven on Wednesday evening - when I might desperately be in need of an inspirational line of poetry, or the reassurance that all is well in someone's household, or a link to a resource or an idea that helps me think through a presentation I'm writing. The signal vs noise distinction often implies a judgement call. The reference Thomas makes is entirely free from this implication -in fact his presentation couches the distinction as personal definitions. I'm just keen to draw attention to the fact that what is signal and what is noise doesn't consist of objective content that we can necessarily pre-determine.

Comments

[Luke Razzell]

Great post, Josie.

I guess we need (time-shifted) search in the mix in addition to timely Rivers of information?

Because, as you say, what you've got may be what I want, but not necessarily right now.

Right now, I'm thinking about how a mix of timely and timeshifted metaphors could work for shared attention around video, with @sleepydog. Hmmmmmm...

[bhc3]

Yes, yes! Great sentiment here. I'm in agreement that the need is not in ever finer controls for filtering. It's in changing our mindset to one of serendipity and discovery. Keeps things fun too.

[Gadget Inspector]

I agree with what you're saying in many respects, however, I've always found with any in depth analysis of social networks, that I find it all too straightforward to be worth such a heitened focus.What I mean is this: isn't (the analysis of social network) all just a bit too obvious and unneccessary?

[Josie Fraser]

Hi and thanks for the comments.
Luke. The bottom line, as bhcs points out, is making things too complex often makes them unusable. So the time issue is more one of digital literacy (on the users side) and better searchability (on the developers/designers side) - ie a metadata or organisational issue.

Gadget Inspector - I always think the aim of good analysis is in part to make the reader think how easy it is, to produce descriptions that seem transparent, so that the reader is left thinking "well, of course, that's completely obvious". I'm not saying I manage it that often - it's a hard job. The value of social network analysis in general? Most users won't and shouldn't have to care. But good analysis can be of use to service providers, workers, and developers, as well as people interested in socio-economic issues around the sector.