tumblr.
quisby.net

19th July 2009

Text

For some time yesterday, I wrangled with the idea of buying Fever. A brief primer for anyone that is seeing that link for the first time: Fever is, in the broadest terms, a self-hosted feed reader that will set you back $30, on top of what you pay for any extra hosting requirements (pro-tip: it won’t work on NFS). Its gimmick — gotta justify a price tag that high somehow — is that links appearing multiple times in your feed get aggregated into one “Hot” list, with the most linked appearing higher up the list. Kind of like a personalised, automated Digg. To wit, quoting Inman himself: “Fever takes the temperature of your slice of the web and shows you what’s hot.”

In principle, the idea is excellent, and interesting step in the domain of feed readers. Anyone that uses a feed reader heavily will know that it’s easy to get overrun by the dreaded “(1000+)”. The majority of feeds just aren’t that useful in most feed readers; the signal to noise ratio is huge. Clearly, a rethink in the way we consume feeds is necessary, and Fever is Shaun Inman’s metaphorical 2 cents, if you will.

That said, I decided it wasn’t worth the price tag. Or even any price, really. (I’m sure it’s a lovely feed reader, with or without the gimmicks, but so are a lot of the free ones.) My reasoning seems to line up a great deal with Alex Payne’s, who, unfortunately, did pay for it. I suggest you read the post if this sort of topic interests you, but I’ll quote the pertinent points:

Fever is just fine for floating good techie content to the top, but poor for most any other subject. I’d love it if Fever could find me good posts from the set of minimal techno or cocktail blogs I subscribe to, but link blogs – and, indeed, linking outside one’s own site – just aren’t as prevalent in those communities.

This quibble alone makes Fever almost useless for me. Granted, Fever seems predominantly aimed at the tech-savvy geeks (it’s self-hosted, remember), so this isn’t necessarily an entirely bad thing, but for those of us capable of installing software on a webhost whose interests are broader than just technology and the web, it’s rather limiting. Be it minimal techno and cocktail blogs or science and language blogs, Fever just doesn’t work outside of a very narrow scope.

Another good point Alex makes regards the pollution of your “sparks” (the high volume feeds from which you want Fever to expose the good stuff). If you’re not careful about what you subscribe to, you’re pretty much just going to get the Same Old Shit; what’s top of Reddit is top of Digg is top of Delicious is not top of my interests.

To expand on that a little: you need to make sure your spark feeds are of a decent quality. But in that case, why aren’t you just reading the feeds directly? Furthermore, the effort that will inevitably go into maintaining your sparks rather belies Fever’s original intent: to lessen the anxiety of pruning and perfecting your list of subscriptions. Same shit, different feed reader.

Alex’s (probably rather more reliable) conclusion is much the same as mine: Fever is an interesting first step in digging through that Information Overload people keep alking about, but it doesn’t nail it, and in fact just introduces a whole new set of problems.

However — and here’s an unexpected twist! — this post isn’t about Fever. It’s about Tumblr.

Fever’s idea does not work in a broader scope that getting the hottest tech news. More succinctly, it just plain doesn’t work in a feed reader. But I think it’s perfect for sites like Tumblr.

Picture this scenario: You went to bed at 11pm, having checked Tumblr one last time. You got up in the morning and went to work (where you don’t have time or access to Tumblr), then you return home at, say, 6pm. That’s 19 hours since you last checked Tumblr. I would estimate that the average heavy Tumblr user follows between 100 and 300 people, which is a lot of posts in 19 hours.

What if, instead of going through page after page after page after page, you could just see the most awesome things from the last 19 hours, and fuck the rest? But how do we know what’s awesome? Hint: the necessary metrics are already in place: likes and reblogs.

Say one of the people you follow posts a link that 3 other people you follow then reblog, and 4 more like. If we use the same point system as Tumblarity does, where a reblog is worth 3 points and a like is worth 1, that post now has 13 points, and would jump to the top of our theoretical “hot list”, over all the other shit that got no notes or just a few paltry likes. Obviously, we only count notes from other people you follow, because you already trust them to post good stuff (you follow them, don’t you?), but they’re like sparks, to borrow Fever’s terminology: they’re unreliable, and potentially high volume — the very nature of Tumblr encourages this. Ideally, you just want the best bits and skip the rest (or view it at your leisure when you have more time).

This does have the unfortunate side-effect of reducing everyone we follow to mere sparks (as opposed to Inman’s preferred “kindling”), but that’s why I think it would be better as a catch-up than a full on dashboard replacement. If you have time, you don’t mind going through the pages of your dashboard. But if you’re in a hurry and just want to see what you missed, you want to eliminate the crap and cut straight to the diamonds.

I think this has an advantage over the standard voting procedures other sites use (like Diggs, and bookmarks on Delicious), because systems like that are inherently unfair: the highest voted links are more visible (because they’re at the top) and thus get a disproportionate amount of extra attention, encouraging more votes, thus further extending their leads, leaving everything in the middle (the same thing happens to the lowest stuff too) to languish. The equivalent “vote” on Tumblr is the already established like or reblog, which as we know, isn’t a vote at all, it’s an acknowledgement of quality. There’s no incentive to get it to the top of some arbitrary list. And if you did want to game the catch up system, you’d be polluting your own blog with things that aren’t really that good (you’d have to reblog to game it, because it’s worth 3 times more than a unseless-in-comparison like).

So on top of it being an excellent way to catch up on the best of what you missed in your dashboard, it’s also a much fairer system of gauging what’s best on Tumblr right now.

Using that logic, you could easily extend this system to every blog on Tumblr and find yourself with a pretty handy zeitgeist-like view of what’s hot on Tumblr. A perfect replacement for Popular, I would posit, because Popular still sucks: it’s a) frequently boring, b) frequently full of unattributed photography, and most importantly, c) only representative of what a few Tumblr employees think is awesome. (Sure, a couple of other people have to like it first to get it in front of said employees, but the employee has the final decision. And I frequently see posts with less than ten notes on Popular, which is not very representative of what the community as a whole is enjoying.) And here we have the same Digg-like problem: what’s popular just gets more popular, what’s in the middle gets no attention. It doesn’t really seem fair.

And I don’t see why this mechanic couldn’t extend to other services that rely on some degree of social input. Twitter, for example, is a good candidate (using the favourites feature to judge the worth of tweets or links within tweets). and Hype Machine already proves the idea can work to a degree — there’s no “community” per sé, but it draws from a pre-defined list of blogs, as though it were a community.

Just sayin’.