LinkRiver Deep Search

The other day my friend Chris shared a cool link on “Barack Obama’s surprisingly non-ideological policy shop”. I follow Chris on LinkRiver, so that link ended up in my river. I read it and shared it, so now the link shows up in my stream and in the the river of those who follow me.

Fast forward a few days and I want to find that link again. Unfortunately, the link title was not very descriptive (“The Audacity of Data”) so I couldn’t use LR’s standard stream or river search to track it down. I could have scrolled through a couple of pages of my stream… but what if I could search the text of the linked story instead of just the title?

Enter LinkRiver deep search. The name’s not great and will probably change, but deep search is a Google Custom Search Engine that makes it easy to search the content of all pages in a user’s stream.

Two caveats. Your first deep search may be a little slow because Google has to grab an XML file from us and build your index before it can return results. Second – I’ve limited this feature to streams (no rivers yet) for now so that the XML file Google has to pull is a reasonable size. I’ll probably add rivers later.

Back to the story. The article I wanted to find was about “obama” and mentioned “policy” and “wonk”. See the results yourself.


LinkRiver Leaderboard

One of these days I will implement a per-stream, per-river leaderboard for LinkRiver showing the top shared links sources. Here are the overall leaders right now:

  26. Most are what you’d expect, but a few stand out. The links generally represent people experimenting with importing their own Twitter feeds into their stream. The and sources represent redirecting links… not the actual destination but an intermediary designed to count clicks. scores the highest because we had a user that tracked the TechMeme firehose feed for a while and generated a ton of links very quickly.

Los Angeles Marathon Pre-Race Info

The LA Marathon makes it hard to find much of the information that racers look for — what’s the easiest way to get to the start and finish, what fuel will be available, etc. Here are a collection of notes, most courtesy of my friend Tod:

Just a little over one week ’til race day!

No Conditional GET for Google Reader, StumbleUpon

Conditional HTTP GETs save bandwidth and computing time by allowing RSS readers to only download feeds when they’ve changed. Most bookmarking/sharing services support this feature (, Ma.gnolia, FeedBurner, NewsGator, etc) but I was surprised to find the neither Google Reader nor StumbleUpon do. This means that every time that LinkRiver checks a Google Reader shared items or StumbleUpon feed for updates, it has to download the entire feed and look for new stuff. Argh 😦

LinkRiver Seeded With A-List Bloggers

Brittney noticed that LinkRiver was seeded with A-list blogger accounts.

So, it seems Scoble didn’t sign up for the account himself, instead he was “seeded,” along with other high-profile Silicon Valley bloggers. This is an interesting marketing move to make.

LinkRiver is a classic scratch-your-own-itch project. There was no good way to aggregate the link blogs published by more and more of the bloggers I read, so I built an app to do it. I added Scoble to the system (Scoble’s stream) because I wanted to follow him, same with many others (see the right side of my river).

Seeding these accounts was not intended to be a “clever marketing move” as much as a practical “I want to follow these people” move. I’ve already had a number of bloggers “claim” their accounts. If it bothers people – I’ll gladly remove them.

LinkRiver Hits The Streets

I’ve been building a link-blog aggregator for a while now. Several bloggers wrote about the service yesterday and have done a better job of explaining it than I could. First Louis Gray:

The newest service to enter the picture is an intriguing entry, LinkRiver, which harnesses your RSS streams from multiple services, including Google Reader shared items, Twitter,, Yahoo! Bookmarks and others, and posts them to a single “Stream”. As your friends join the service, or you choose to subscribed to other LinkRiver users, these small streams become a “River” of shared links, hence the name.

More from Frederic of The Last PodCast:

LinkRiver takes its inspiration from services like Friendfeed, ReadBurner, as well as the metaphor of following from Twitter, and mixes it up into a cocktail that looks deceptively simple, but is utterly addictive. The interface is extremely clean and very fast.

More to come!