Archive for the 'Arrington' Category

Loving aideRSS

Tough love, that is—there’s a lot more I want out of this.

But first, aideRSS is awesome. When I serve it a blog’s feed, it looks at how many comments, delicious saves, and other mentions each post has and then divides them up according to their popularity relative to one another. AideRSS offers me a feed for each division—the smallest circle of the “best posts,” a larger circle of “great posts,” and an even larger circle of “good posts.”

I’ve got two main uses for it. It ups the signal-to-noise ratio on blogs that aren’t worth reading in their filtered state, given my peculiar tastes. And it allows me to keep current with the most popular posts of blogs I don’t have time to read every single day. That’s huge.

There are real problems, however, and other curious behaviors.

Consider Marc Andreessen’s blog pmarca. For one, AideRSS strips out his byline (here’s the “good” feed). For two, it has recently really oddly clipped his most recent posts and made them partial feeds (I also follow Andreessen’s full feed, and it is still full). Also, aideRSS also seems to strip out all the original dates and replace them with some date of its own.

That’s a problem. Google Reader published Andreessen’s post called “Fun with Hedge Funds: Catfight!” on August 16, 2007. But it’s the most recent post in AideRSS’s filtered feed of Andreessen’s “good” posts. The problem is that it follows “The Pmarca Guide to Startups, part 8” in the “good” feed but precedes it in the regular feed.

Did the post about the hedge funds and the cat fight receive some very recent comments, more than a few days after it was first published? All else equal, it wouldn’t be a problem to have the posts out of order—that would seem to be the sometimes inevitable result of late-coming comments or delayed delicious saves, etc. But all else is not equal—because the original dates are stripped. Posts in a blog exist relative to one another in time. Stripping out the dates and then reordering the posts smothers those important relationships.

But let’s look to the horizon. AideRSS can’t handle amalgamated feeds. I want to serve it what Scoble calls his link blog—the feed of all the very many items he shares in Google Reader—and receive only the most popular. That way, I would get the benefit of two different kinds of networked news at once. I’d get the intersection of the crowd’s opinion and the trusted expert’s opinion.

I’d also like to serve it a big mashup of lots of feeds—say, my favorite five hundred, routed through Pipes—and have it return the top two percent of all posts. That kind of service could compete with Techmeme, but it could be dynamic. We could all build our own personalized versions of Techmeme. That would be huge.

Trying it out a few different ways gave wild results. The posts in an amalgamated feed weren’t exactly being compared to one another on a level playing field—so that even a relatively bad TechCrunch post with ten comments crushes an small-time blogger’s amazing post with eight comments. But they also weren’t being compared to one another only by way of their numerical rankings derived from their first being compared to the other posts in their original feed.

Why can’t aideRSS measure each post’s popularity with respect to its kin even when it’s among strangers? The share function within Google Reader gives aideRSS the original url for each post. Can’t aideRSS take the original url for each post, find the original feed for each post, and then analyze each post against the other posts in its original feed? That would be much more analysis, for sure, but it would also be much more valuable. I’d love to see it.

Of course, while it may be a surprise or unintuitive at first, all this is really just one particular take on the first and second components of networked news—pulling in your news from a network of publishers and from a network of readers, including friends and experts and others. Without my additions, aideRSS represents just the second component, in which we get news based on whether others are reading it and participating in the conversation around it. My additions bring a little of the first component.

UPDATE: It would also be awesome to serve aideRSS the feed generated by a WordPress tag or by a persistent Google News search. That would be bringing in a shade of the third component of networked news.

Advertisement

Applying the Three Components of Networked News

For a goodly time now, expect to see this blog try to flesh out this concept of networked news by reviewing isites that concern themselves with the news, either completely or substantially—from the New York Times to digg to Google Reader to Mario Romero’s awesome Google Reader Shared Items app for Facebook to Topix to Memeorandum to Pageflakes to Thoof to Streamy to whatever else with which I may cross paths.

I’ve got my three metrics, and it’s time to make sense of them. Maybe along the way we’ll figure out what some bright person could do to satisfy Arrington’s appropriately underwhelmed feeling about online news. Maybe we’ll figure out that “networked journalism” has something to do with networked news. Maybe not.

But what’s first?

Ornery Arrington on Personalized News

streamyI can’t get enough of how Michael Arrington, at TechCrunch, rags on personalized new sites. I love it. Talking about the newest venture to join the hunt for a killer solution to online news, Arrington writes of Streamy, “It’s pretty and extremely well thought-out, but it’s not clear that it does anything new enough to grab people’s attention.” Plus, “It is well designed, has lots of intelligent features, and is almost sure to drop into obscurity immediately after launch.” What a guy!

For now, I’ll withhold real judgment—especially of the claims made here—till I’ve had a chance to check out the private beta, to which I’ve requested an invitation.

Meantime, I’m impressed with the social networking and what they call “filters,” which are essentially substance- and source-based ways browse, and subscribe to, kinds of content, by keyword and original author, respectively. Ideally, that would mean that someone could set up his personalized page to pull in everything Arrington says about, say, “personalized news.” The trick, then, is how social networking brings the virus to the filter. Your friend’s becomes your filter and then it becomes my filter; my filter becomes yours, your friend’s, his friend’s, and so on.

Note that amateurs don’t have to control these filters either. Arrington could promote his own, using them to help him slice up his content along different lines, potentially overlapping lines, and push it out Streamy readers of news, if there ever are any. Or these filters could allow someone to become an editor, much the way Scoble acts as an editor (choosing the news) with what he calls his link blog. If I were running Streamy, I’d be thinking about how I could allow users to monetize their filters. They are, after all, just platforms for subscriptions. Find a way to allow an expert on some tricky topic like global warming sell you his daily digest of the best reading on climate change. He could write his own original material as well, of course, and include it in his filter.

I agree with Arrington that integrating IM is smart, and integrating the ability to drag and drop stories is super smart. I wonder how popular it will become for sites to integrate IM. Will the New York Times have it some day? Will espn.com? It seems to make sense to distribute IM over virtual locations rather than keep it all cooped up in one place. If I were running meebo, I’d be thinking about how I could build a proprietary web-based IM client just for the New York Times. (From what I can tell, Jake Jarvis has tried to turn meebo into a Facebook app. I signed up, but got an error: “There are still a few kinks Facebook and the makers of Meebo are trying to iron out….”) I love meebo almost as much as Arrington’s orneriness.

Here’s the screencast of Streamy:

Thoof Is Close But Not Nearly Close Enough

ThoofI think recently launched online news site thoof has taken another step in the right direction—toward giving its readers the news they want and withholding the news they don’t. Thoof seems to learn what its readers like buy remembering their clicks on article summaries that given user-assigned and -edited tags. If you click on an article summary with the tag “politics,” the next time you visit you’re more likely to see article summaries tagged “politics.”

But Mike Arrington’s concern, inspired the fact that “web is littered with failed or stagnant personalized news startups,” is still relevant. People do “usually want to read news and then discuss it with friends. That is why people still tend to prefer big news sites: “everyone else flocks there, too.”

If it wanted to, however, Thoof could keep something like main page that displayed the aggregate of all its readers’ preferences. Each reader could then, at any time, choose whether to read his personal page or the page that reflects the many, many decisions of the Thoof community. Then Thoof might be able to satisfy Arrington’s worry that “the niche audiences that really want personalized news aren’t enough to sustain these startups.” Thoof can have it both ways.

Better, in fact. Readers who want the news the “flock” sees get a much purer version of it, an elevated version of that. The flock isn’t just reading the same thing—they’re causing themselves to read the same thing along the way. They’re not just readers, led by a shepherd. They’re editors, and they lead themselves.


Josh Young's Facebook profile

What I'm saving.

RSS What I’m reading.

  • An error has occurred; the feed is probably down. Try again later.