Google’s Summer Of Code (featuring OSSstock)

Google just announced its Summer of Code initiative, inviting students to build open source projects for a $4500 reward. Every team that Google accepts to work on a project gets a $500 stipend, and whoever successfully completes their project gets $4000. In addition, each group gets a mentoring organization (like The Perl Foundation, The Apache Software Foundation, The Wine Project or Google itself), which also gets $500 (unless its Google; they get nothing).

From the FAQ:

What kind of deadlines are associated with the Summer of Code?
    May 31st: Announce on code.google.com and open application process

    Interim Period: Back and Forth with applicants on Summer-Discuss

    June 14th: Final applications approved. Cut $500 checks for initial funding.

    Interim Period: Mentoring organizations give the students a helping hand, guidance.

    August 3rd: Google gives a preliminary progress report at OSCON.

    September 1st: Deadline for all the student work (pencils down).

    September 30th: All adviser feedback in.

    October 1st: Announce successful participants. Cut final checks and send t-shirts. Party?

Each mentoring organization has a list of ideas. One of them involves fixing “Awful hacks“, or
Identify and eliminate inelegant solutions which were implemented due to time/release pressure, and replace them with proper solutions.
as well as creating artwork, porting Windows applications, doing PR work, writing a FAQ, or about a billion other things. Not only will this help the open source organizations involved (the promise of some cash can be a great motivator), but Google is helping to identify some talented programmers it might want to hire. And who said Google didn’t give back to the open source community?
(via Search Engine Watch)

UPDATE: Nicholas let me know about the Slashdot article announcing the project. The article is submitted by Google’s Chris DiBona, their Open Source Program manager. Chris also sticks around to address any questions in the comments. See, Google, Chris (a former Slashdot editor) understands all the things about blogging that the Google Blog gets wrong.

Posted:
May 31, 2005 by Nathan Weinberg in:

How Hard Is Searching?

Gary Price makes an excellent point about all the various ways a good search engine can still yield no useful results for some searchers because of thier behavior, including the tendency to not look past the first few results and the amount of effort and time invested. I know many people who consider themselves “power searchers”, myself included. If you ever wanted to know why Google hasn’t conquered search yet, the fact that we need that term tells you everything. Any future search engines that let non-”power users” find things based on how they search is going to topple the market leaders, just like Google did.

Posted: by Nathan Weinberg in:

Yahoo’s Guide For RSS Publishers

Yahoo has put together a handy guide for RSS publishers, dealing mostly with how to produce a feed that will work well with Yahoo’s services. Their “Publisher’s Guide To RSS” is a must for anyone who publishes through RSS, considering the considerable clout of the Yahoo network. It includes:

  • An explanation of what RSS is for those who are new to the concept (Yahoo advises RSS 2.0, although it supports all versions and Atom)
  • Links to tutorials on creating feeds and using them
  • Typical link structure for hosted blogs, such as Blogger, LiveJournal and TypePad
  • How to validate your feed
  • A box to submit your feed to Yahoo, so those searching for content in MyYahoo (and a possible future feed search engine) will be able to find them (Yahoo requests you only submit one version, preferably RSS 2.0, to avoid duplicates).
  • How to ping Yahoo when your content is updated
  • How to add metadata to your feed so it ranks well in their feed search and displays descriptions in feed readers
  • How to request your feed for inclusion in the Yahoo RSS directory
  • Simple code for an “Add to My Yahoo!” button
  • How to enable RSS autodiscovery for compatible browsers and toolbars
  • How Yahoo leaves subscriber data in your site logs

There plenty of stuff in there, and its all short and consise. Take the 15 minutes to go through it; its worth it.
(via Search Engine Watch)

Posted: by Nathan Weinberg in:

Yahoo Music Hacked

Robert Chapin says that his company, Chapin Information Services, has found a flaw in the design of Yahoo’s website that allows easy, free downloading of DRM-free music tracks. Of course, Chapin isn’t saying how to pull it off. Chapin says the flaw is simple enough, so simple in fact, that it is easier to download the free, stolen track than to pay for it. Making this more useful, the stolen music is not truly from the Yahoo Music Unlimited service, but rather the music that costs 79 cents per song and comes without any copy protection or restrictions.

P2Pnet says they received an email from Yahoo where Yahoo indicated that after seeing what the problem is, they have determined that it does not expose any sensitive user information, just free music, so that it is not serious, although they are looking into how to fix it. Looks like Yahoo has its priorities in the right place.

More:
Om Malik
Engadget

Posted: by Nathan Weinberg in:

Findory Adds Ads, Personalizing Google AdSense

Findory has done the impossible, personalizing Google ads to the user on its site, presenting surfers with mostly high relevance ads on all its pages. Greg Linden announces it.

First off, a quote from Greg’s blog:

This early version is built on top of Google AdSense, but these are not normal AdSense ads. They are not targeted merely to the content of the page, but to the individual behavior of each reader.
This is a very interesting statement, particularly where he says “built on”. My assumption is that Findory has somehow worked out a system that integrates its user data with Google’s ad network to generate much tighter ad matching than Google’s page reading contextualizing algorithms. Of course, he could mean something else entirely, so I’m waiting on more info from Greg. Greg would only say that it involved some work with Google, but most of the work is done on Findory’s end that allows Findory to better target the ads.

How good is the system? Very good. Greg actually uses the Findory source page for this very blog as an example. The ads I’ve seen are related to search engines, PageRank, newspapers (a recent post), blogging, RSS, XML, Newsgator, enterprise software, AdSense, AdWords, affiliates, portals, although they sometimes miss, such as when I saw a bunch of ads for cars or freight trucks, or clip art. The Slashdot page saw the world’s geekiest ads. The reason these ads were useful? Most of the keywords in the ads appeared nowhere on the page, but where completely relevant to my blog. I can only reason that Findory is passing along accurate data to AdSense about the source on the page, allowing for a variety of useful ads.

Now, while the source page ads can be generated irrespective of who is viewing the page, the Findory home page ads cannot. Every Findory home page is customized, even if no one is logged in, and has no unique identifier. It is impossible for Google’s bots to see my Findory home page, so the data being fed to personalize them must be coming through a different channel.

The ads on my personalized Findory page detail Google Local, popup blockers, search toolbars, RSS, MSN, music purchased online, Internet Explorer, antivirus, Google Desktop Search, XML, SOAP, news tickers, Windows, the stock market, email, spyware, left-wing attack ads, Launchcast, radio, surveys, web cams, blogs, source code, IT, spam, the legal system, the pope, audio software.

What did you not see in that list? Not a single ad for a physical product. No ads for cars, books, cameras, food, porn, even electronics. Why? Because I am a non-commercial surfer. I use Findory a lot, but I am never reading about products or gear or clothes. Even though articles about those topics appear on the page, the ads only reflected what I read. Notice what else isn’t there? Celebrity based ads. Nothing about Britney, or Desperate Housewives, or Rebecca Romjin. And those certainly appear in the non-personalized news stories. However, I don’t read those stories, and Findory doesn’t have any information that says that I do.

How does it work? It appears Findory is feeding Google keywords through the keyword suggest tool, keywords that are altered on every load of the page. In the page source code, I see this:

google_kw = "Tech about blog believe here three source been radio open checkpoint because police monday running currently support john spam simon";
Reload, it changes:
google_kw = "just Tech radio believe talk about listening being child General kiss another blogs record explorer treasure records hour been format";
Reload:
google_kw = "Tech just about talk believe radio says interviews treasure prague major trustrank Science beta apprentice said Personal hours newspapers around";
Reload:
google_kw = "Tech just radio believe talk about glitch newspapers trackback podcast length format times around explorer listening loves extra riffing record";
You get the idea. What does Findory do to get those keywords in there, changing on every page load, different for every single use? Is it a hack, a brilliant script, a deal with Google, or something else? Either way, its very effective, and I wish I could use it on my site. Its very similar to how I have WordPress automatically insert a post summary in the (still malfunctioning) Y!Q code on every post page, just far, far, far more advanced. Great way to make sure AdSense actually knows what the hell is going on, since it tends to make a lot of mistakes.

Oh, and congratulations to Findory. I’ve always loved the site, and it is good to see they are at least going to start seeing some revenues. Interestingly, Findory needs these mega-targetted ads more than most. Its site finds such accurate content, that if the ads were not equally accurate, they would be so below the level of the page content as to ensure that no one ever clicked on them. Now that I know that the ads are pretty accurate, I’ll be paying close attention to them (and hopefully send a little coin their way).

Posted:
by Nathan Weinberg in: