Skip to main content

Platform Dynamics is a huge but underappreciated, and difficult-to-articulate element of what makes online spaces...

Originally shared by Edward Morbius

Platform Dynamics is a huge but underappreciated, and difficult-to-articulate element of what makes online spaces work or fail, as well as defines the types of engagement they produce.

(Elevated from Dran Fren's excellent thread here: https://plus.google.com/+DranFren/posts/VAy7fArGAwE)


For Google+ specifically:

The "subscribe to post" dynamic when commenting.

The unlimited time period for post activity (they don't close after a day, or week, month, year, etc.)

A post owner who practices strong post hygiene (removing crap and low-level comments).

And the G+ Notifications Pane (during various of its Better Behaved incarnations, the present era being one of those)

... means that a good discussion can take off and live over a long period and is worth returning to.

Mind: these are warrens within a warren-nest, not plazas, and good and highly-productive plazas on G+ are hard to find, organise, promote, and maintain.

(This ... is not dissimilar to other platforms, mind.)

But: The G+ dynamic is highly unusual amongst other platforms.

On Reddit, additional comments to posts do not alert any participants other than the direct parent. Reddit is where discussion goes to die. It's really sad, because it's otherwise a powerful platform.

The fact that large Reddit discussions would be intolerable under the G+ dynamic does not mean that this couldn't be applied to small ones.

On Hacker News, discussions only "live" within visible sight for a few hours, possibly a day or two. Conversation rapidly ages off the front page. There is no notification dynamic. Readers can check recent activity for responses but aren't otherwise notified. This keeps flamewars down, but also kills long-lived discussion. HN's moderation, reasonably selective readership, topic curation, and threaded response do allow for high-quality medium-lived discussion (a few hours, a day or two at most), but not more than that.

Tildes tends to share HN and Reddit dynamics. I've found it hard to gain traction and to sustain interest there.

Ello actually improves on Google+'s dynamics (subscribing to threads is independent of commenting), but the small size and other frustrations of the platform ultimately killed interest. There was a design phase where the Notifications panel was a sidebar (left of screen), and the currently active notification item opened in the main pane. I am so much in love with this model. You can skip through Notifications, see full context, read/respond, then move on. Closest I've seen otherwise is the 2-3 pane Usenet newsreader display, with active posts and thread view active (Netscape 3/4 Usenet client, rtin, etc.)

A surprisingly high-impact decision at Ello was to collapse the following dynamic from two groups to one. Previously there'd been "starred" and "noise", later renamed. This created the option to have high-interest and low-interest streams, and while having _only two streams was limiting, collapsing this to one proved fatal. Many people (myself included) responded by unfollowing all of their "noise" contacts. Even if those had been secondary, they provided a useful mixing and extension layer between primary and other contacts.

Imzy had a stunningly bad dynamic resulting from numerous interactions, non on their face obviously bad.

The site had a strong focus on anonymity and pseudonymity. Users and/or discussion forums could opt to have single-instance anonymous identities assigned for the duration of a single post.

Blocking of anonymous identities was not permitted because that might leak identity information. (Block an anon, see what non-anonymous user was blocked, unmask anon.)

Notifications could effectively not be turned off.

Notifs didn't clearly indicate what or who was triggering it (other than the discussion), and the semi-threaded style was ... tremendously confusing to read. So with a new notification, you'd have to wade through a post trying to sort out who'd just triggered it.

The founding cohort (quite probably legitimately) felt disempowered elsewhere. A substantial and extraordinarily difficult-to-ignore contingent took it upon themselves to use their power at Imzy to harass those they took (or otherwise determined to or labeled) as harassers. Much as the Wil Wheaton / Mastodon incident reprised this past July-August.

Forum moderators were AWOL, staff were entirely ineffective and denied the problems. The fact that there were so many individual discussion spaces amplified the problem (more corners for problems to arise w/o adult supervision).

The result was mind-warpingly disturbing. Even knowing it really didn't matter, it was hard to distance myself from it. And best I can tell, Imzy's founders and staff went in with good intentions. They were trying to solve the wrong problems, or even possibly, the right problems (Reddit at 10 milliion+ users) at the wrong scale (Imzy at ~10k users).

SlateStarCodex describes the Toxoplasma of Rage specifically in the context of Tumblr. Well worth looking up and reading. Sites optimising for engagement can be highly problematic.

danah boyd has written much on 4chan and related culture. Recommended.

MetaFilter has an outstanding S/N ratio. Something that turned up in my earlier analysis. Tiny site, but very well curated. Constantly on the financial edge, users effectively bought it back and finance it now. Somewhat stale, but viable. See also The Well. Both are Old School Online Communities, the type Howard Rheingold had in mind (and direct experience with) when he started writing about these communities. Their self-selection and soft-but-effective barriers to entry make their experiences non-generalisable.

Comments

”go"