5 Comments

Summary:

GigaOM technical architect Nitin Borwankar usually stays behind the scenes, but he got so excited about vertical code search he stepped in and wrote a post about it. STARTUP: Krugle ELEVATOR PITCH: “Find, fix, and learn.” Vertical search engine and web-based work support for developers. WHAT […]

GigaOM technical architect Nitin Borwankar usually stays behind the scenes, but he got so excited about vertical code search he stepped in and wrote a post about it.

STARTUP: Krugle

ELEVATOR PITCH: “Find, fix, and learn.” Vertical search engine and web-based work support for developers.

WHAT THEY DO: Create value for developers who use open source and freely available software, by reducing pain of search and enabling re-use. Includes 500+ public repositories indexed and cached for performance and 50,000+ open source projects. Also provides tools for saving the state of a workspace (multiple open code files) and then sharing the URL to the saved workspace, a very powerful collaboration tool (see example of some code for creating CAPTCHAs).

BUSINESS MODEL: Contextual ads, affiliate programs (e.g. O’Reilly’s Safari), memberships.

SECRET SAUCE: Highly disruptive open source search engine “Nutch.”

THE DEAL: Ken Krugler, Krugle CTO and founder, likes to ask, “what is the most commonly used developer tool?” Most developers might think it’s their code editor or IDE. But Krugler is right — it’s search. Developers spend an inordinate amount of time on the net searching for code, information, answers and explanations for cryptic error messages. And they do it over and over, in a highly fragmented and unproductive ways.

As the Krugle team tells it, there are vertical search engines have three major characteristics that are different from horizontal search engines.

  • Specific information. Information is more constrained and has unique characteristics depending on the vertical. Often multiple types of objects are being searched over, and results collated.
  • Traditional ways of determining relevance don’t apply. Counting links to a web page isn’t enough inside a vertical domain, some semantics need to be taken into account
  • Traditional UI is not enough. Because keyword search is not enough, the UI needs to support exploratory search.

There have been some relatively cursory reviews of Krugle vs Google Code vs Koders, but these miss the point that raw search is not all that Krugle serves up. For instance, when I searched for “MD5″ as the above reviewer did, I also looked at the ‘tech pages’ tab and the ‘projects’ tab. Together with the code tab there was far more useful information than when I did the same on Google Code, which had just a bunch of header file with the text md5 in them, and at http://www.koders.com, where I did get some project results on the side, but when I dug deeper I found the UI far less conducive to workflow.

One nit is, while the center part of the home page has a lot of Ajaxy UI goodness, the actual meat — the search field — is hidden away at the top left. People are used to seeing a search field right in front of them if they are going to a search engine. However, the customizable Firefox plug-in is very useful.

I wish more people knew about this site and used it — we would all be far more productive as developers by not having to find and re-find the same open source projects and web pages over and over again.

You’re subscribed! If you like, you can update your settings

  1. Doesnt sound that exciting. Sorry. Flop.

  2. Well, that was a curt pan, Diana…

    Myself, I’m not a developer, but I am in the software industry. (Okay, I admit it, I’m a marketer!) I think Krugle sounds like it’s really onto something. I’m sending this post to the developers I work with.

    [In the name of transparency, another admission: I know Krugle people and like 'em. I think they're a very smart bunch, and they're working really hard to change the world.]

    Net-net, I think if the architect of GigaOM likes Krugle, people should pay attention…

  3. Hi Diana, Graeme,

    Couple of points.

    One – in any vertical category if you are familiar with the domain it’s going to be interesting, if not the ‘ho hum’ neuron is going to kick in.

    Two – the hidden bigger point here is Nutch, the open source search engine. This is looking like it will be the Linux of search – making the power to create search engines available to a much wider audience of developers.

    Just when it seemed the options were between one mainframe and the other along came the mini’s, then the desktop machines. The pendulum has continued to swing between the two architectural ends.

    To me Google, Yahoo, MSN are the mainframes of search. The desktop of search is something like Spotlight (on the Mac). So what’s the mini of search? My gut feeling is Nutch will be powering a whole family of vertical search engines.

    Moreover, Doug Cutting, the creator of Nutch, is employed by Yahoo who is reputed to be dedicating serious resources to Nutch. Read between the lines …

    You heard it here first ;-)

  4. Great post, Nitin! [Dang, I need to hear that phrase more often. :-)] I think you’ve hit the nail on the head – the big advantage vertical search engines like Krugle have over the large horizontal search engines is in the additional data and related services they provide above and beyond raw search.

    Google itself has recently been making forays into vertical search functionality, with Google Coop, OneBox and custom search engines. (The mainframes are fighting back!). Vertical search leaders aren’t fazed by these changes, though!

  5. Confluence: Marketing Monday, December 4, 2006

    Press Hits…

    Dec 4th 2006 Bad News Sells…

Comments have been disabled for this post