Thanks for the comment, and I agree, it's very bare-bones right now.
In an ideal world, what would a product like this look like when fully fleshed-out?
In an ideal world, what would a product like this look like when fully fleshed-out?
There are a lot of sites that I participate in that allow you to rate and flag other users (e.g. HN). I actually really like this ability. But frequently I want to make many posts to such site, but I don't really want all my comments to be able to be tied together
For example, there's an HN post about BFE, Idaho and I mention that's my hometown. In another post, it's relevant to the conversation, so I mention I'm Italian. In another post it becomes clear that I have extensive professional experience in underwater basket weaving. Put these three facts together and they can quickly undermine any anonymity (i.e. I'm the only Italian born in BFE, Idaho that has a career in underwater basket weaving).
So having a log-on system that allows me to be held accountable (i.e. If I make a comment on HN that gets downvoted, my HN rating still goes down) but where I can still be anonymous (i.e. no one can aggregate my comments). From a service perspective, I'm thinking of something like a Disqus plug-in architecture where I could log-in to the primary service, and only that service would ever know that I'm a singular entity. That service would provide an api that other sites could use set and grab general reputation data.
For example, I log into Service as 'yaddayadda'. I navigate to HN, and log-in. The Service reports to HN that I'm an authorized user currently known as 2051013165142 (new, random user id each time I visit HN). HN polls the Service API for my current user id. The Service reports back that my identity has earned a total of X HN submission up votes, Y HN comment up votes, and Z HN downvotes. HN can then do what it does to calculate my reputation. When I subsequently make a submission or comment and am rated, HN reports those points back to the Service, which aggregates them into 'yaddayadda's' HN totals.
I have played with gender-identification software[1], which has such a limit. Frankly, the identification seems to be more topically sensitive than it is gender sensitive. For example, if I put in my writing about robots it will id me as male, but if I put in my writing about cooking it will id me as female.
Combining the 'author identification' software with known identities (such as we have here on HN) will certainly make them more accurate. Hence, my desire for a system that limits such computations.
That said, it might become a more attractive experiment if it had some way of floating good thoughts to more readers. E.g. you could "like" only one of the thoughts from your feed and have an extra slot in your feed for a thought that another user "liked."
We're on the same page there. Still it would be nice to find a way to make enough to cover the server costs without fucking up the site.
The notion of upvoting thoughts is a recurring theme in this discussion, so I'm bumping that to the top of the todo list. Again, thanks for the input, it's very much appreciated.