A couple of weeks ago (mid-June) I’ve participated in our internal Yahoo! EU Hack Day. This time around I’ve decided to tinker with Yahoo! SearchMonkey by building a couple of modules (dubbed SearchMonkey apps) during the 24 hours available.

In case you have never heard about SearchMonkey: it’s a technology built into Yahoo! Search that allows developers to present their own or other peoples’ sites in a different way (examples to follow). These search result augmentations are dubbed “SearchMonkey apps”.

The Search users then have the option to use the SearchMonkey apps they like, which enhances their search results by presenting them in a different way; review sites for example might present their articles in a way that includes the rating or verdict of the reviewed item. There is a good number of different apps available at the SearchMonkey Apps Gallery.

Using SearchMonkey doesn’t influence the ranking or the order of the results; it displays the results in the same order, just differently.

Anyways. I ended up writing 6 apps, and most of them I’d like to share. For more information about a particular app, just click the links. They’ll take you to the related Gallery listing, where you’ll have the option to enable the app for your searches.

Twitter Profiles

Whenever a Twitter profile turns up in your search, you’ll get a bit more information about the person—number of followers, number of people followed, name, bio, location, last tweet. like this:

By the way, it doesn’t query the Twitter API; instead it just grabs the HTML page and extracts the publicly available information.

Once you’ve visited the Gallery page and installed it, try a Yahoo! UK example query to see it in action.

(On a related note, there’s apparently another Twitter SearchMonkey app, but I couldn’t find the installation link to give it a try. Bart, the developer, does things in a more complex and probably more fail-safe way; in my defense, I wanted to keep it simple.)

Wikipedia Quick Lookup

If a Wikipedia article is found, the app will show you its first paragraph and, if available, a thumbnail of the first image. Also the category:

Once you’ve visited the Gallery page and installed it, try a Yahoo! UK example query to see it in action.

Xing Profiles

Xing is a more Europe-centric LinkedIn contender. It’s big in Germany. If a Xing person is found, the app shows the mugshot, the location, the title and the industry. Unfortunately Xing doesn’t reveal the company on public profiles unless you’re logged in, which my SM app clearly is not. Still:

Once you’ve visited the Gallery page and installed it, try a Yahoo! UK example query to see it in action.

Vox Profiles

In a desperate attempt to suck up to Anil Dash (I kid) I’ve hacked together a module to show the profile of any Vox blogger when a related page is found:

Once you’ve visited the Gallery page and installed it, try a Yahoo! UK example query to see it in action.

TechCrunch / Valleywag summaries

Using cutting edge high tech web 3.0 algorithms straight out of Munich, Germany, this module is able to extract the gist of any given Techcrunch or Valleywag article.

I don’t like these sites very much. At least their Yahoo! coverage (which I am interested in, given my background) is usually hysterical and often times sensationalist. DO. NOT. WANT.

Once you’ve visited the Gallery page and installed it, try a Yahoo! UK example query to see it in action.

Well, that’s it. Maybe you like those, maybe not. :) If you’re interested in writing your own apps for Yahoo! SearchMonkey now, take a look at the official Manual for SearchMonkey Developers and Publishers! And if you’re now wondering what other applications might be out there, the Gallery is a good place to start.

UPDATE/DISCLAIMER: Even though I work for Yahoo! during the day, these are NOT official SearchMonkey apps by Yahoo!. You’re looking at personal stuff. Again, this post, as everything on this site and under this domain, is NOT YAHOO! BUSINESS. Thank you.

Remember the other day, when I told you about the user reputation features in Twerp Scan? Yeah, good times.

Anyways, today I’ve cobbled together a quick Greasemonkey script to display the reputation and voting thumbs on Twitter user pages, right above the “About” portion in the sidebar. Like so:

So this means that you can see on a glance whether someone is naughty or nice, right on his or her user page. And you can vote right there, on the spot, if you feel like, and this will of course feed the same database as you rating your contacts in Twerp Scan.

Use Greasemonkey for Firefox or GreaseKit for Safari to install the script, and then visit any Twitter user page.

Another late night Twerp Scan coding session. I mean, why not? ;) So, what’s new today?

• User reputation: You know the people you’re following, right? If you like what they do, give them a thumbs up or down rating. (No registration necessary. Just click.)
• Some more pagination options: If you want, you can see all your contacts at once now.
• The table columns can be re-ordered. Want the verdict right next to the name? Click the column header, hold the mouse button and drag it over.
• More vertical space. It’s a wide table, after all.

For me, the reputation is a big one. Up until now, there was no way (that I know of) that allowed for community-driven behaviour rating. Here’s the idea: we can block the “bad guys” all we want, but if I’d block @buyinsurancelol, you wouldn’t know. But now I can give the guy a “thumbs down” in Twerp Scan, and if you have him in your contact list, you will see he’s got a thumbs down.

On the other hand, if a 100 people think @MyBlogLog is an okay guy (he is) and give his Twerp Scan contact list entry a “thumbs up”, you’ll see it in your contact list (on Twerp Scan)—which might convince you his 1:7 ratio and his following 15k other people are okay.

Well, at least that’s the idea. It’s a community-driven experiment. So consider playing around with it for the betterment of interwebs! Since it’s a pluggable system I think there are other possibilities… For example a Greasemonkey script that adds the reputation thumbs to the actual Twitter pages etc. Let’s brainstorm. The comments are open.

Well, that’s it. Head on over to Twerp Scan to check out the new features. And please let me know what you think, either via Twitter or the comments below.

(By the way, big thanks to the YUI team, the Yahoo! Pipes and JS-Kit for making this endeavour so simple. Appreciate it! Now if there only was something so simple to make the page less fugly…)

Evening, y’all. I’ve been working a bit on Twitter Twerp Scan. Here’s what’s new:

• It should be much faster now, due to changes in the processing logic. Up until now, the table was rendered, and then updated once for every scanned follower. Highly inefficient. Now the data is fully loaded before the table is rendered, which helps (at least my) processors to not go up in flames anymore. Let me know how it works for you.
• There’s a nice loading indicator now, and a readable countdown! This is the best day of my life.
• Table pagination should help readability: by default, the table is now chopped into bite-sized 25 rows per page. There’s a handy little dropdown where you can change this setting.
• Bugfix: Not broken in Flock anymore.
• Bugfix: countdown accuracy.
• Again, there’s a Twitter account for the highly exciting TS minutiae now, @twerpscan. You might want to keep an eye on that one if you’re interested in the service.

Twerp Scan received a good amount of hits over the last few days, mainly due to a post on Download Squad. Honestly, I am a bit surprised about this, but I won’t complain. Nice to see my little experiment is filling a niché for some people. :)

Let me know what you think of the new features. I have more things on my todo list, but my day only has so many hours (and I have a day job).

Good night!

Over the last few days I’ve pursued my idea to build a “Twitter spammer detector” of sorts using only client-side technologies. I didn’t feel like setting up server components at all.

What was it supposed to do? It should check the number of followers of everyone on your contact list, the number of people they are following, and the ratio between those. If the person is following more than 1000 people (can be customised), and has a Following-to-Followers ratio higher than 1:1 (can be customised), you’d be informed (by a handy “Block” link).

Thanks to Pipes, the YUI libraries and the Twitter API returning JSON, I was able to finish the first version of Twitter Twerp Scan in a relatively short period of time.

My Javascript-fu is still a bit weak, but getting better. It’s a nice change to only have one local file to work with (excluding CSS, of course).

Over the weekend, the traffic to that page increased quite a bit, mostly because several people liked the idea enough to link the site.

I’ve set up a Twitter account for service updates etc. — twerpscan. The rest of my tweets will go to Carlo as usual.