@alice i take the view that "to my contacts" is the kind of thing you would publish on your blog etc for. this is inherently different than having a discussion about nothing in particular

@tech_himbo sure but where do you start? like a new person showing up without knowing anyone. this is the equivalent of like "i just moved to a new city and i know zero people there" -- how do you meet your first friend? your first 10 friends?

@trwnh this is where the web has an edge. most “getting started” UIs ask users for a list of interests, then present a few topics/accounts to follow. the closest real-world analogy for this is hanging out at a place where you know the crowd has something in common. but in the real world, you have to do this by trial-and-error, because you don’t really know what people in a crowd like without talking to them. online, you can learn someone’s interests via the stuff they post and interact with

@tech_himbo that's true but i'm more wondering about the contexts in which you might discover existence (before digging deeper). the only answer i can come up with is something like directories, which is a bit indirect in that you have to crawl a bunch of links which can be overwhelming

@trwnh by “discover existence,” do you mean the existence of a specific person you’d like to talk to? (as opposed to a good post or a neat forum)

@tech_himbo discover a blog, a forum, a website, whatever -- there is an eventual path to "subscribe to see more publications from this person" or "connect with this person to message them" or whatever. but you gotta start somewhere on the Social Web, and i think we need to have a better answer than just "directories"

@tech_himbo like making friends might be the end goal here but to do that you have to meet people

@trwnh the evolution of this over time was:
- directory (stream of raw data)
- feed aggregator (combining streams)
- hand-written filtering rules (e.g. “posts tagged ‘music’”)
- ranking based on crowdsourced data (reviews, upvotes, etc)
- recommendation engines

each step gives a view on a larger data set, with more advanced filtering and relevance ranking. this is needed to scale. imo, if we reinvent directories today, we’ll reinvent recommendation engines tomorrow

@tech_himbo i think that would probably boil down to having an "indexer" as performed by a "discovery service" (especially if done at the "post" level) but i'd still like to improve the model further

i’m pretty sure the most useful thing about twitterlikes is that they facilitate triadic closure — if A talks with B, and A talks with C, then there is a higher chance that B and C will start talking with each other

i’m just wondering if there’s more to it and if there are other models that do this comparably well

i guess networking events? like job fairs but for making friends

the key is knowing at least one person and being comfortable with them before branching out to mutual connections, but getting to know that first person when you have zero connections is hard