I asked Steve Larsen, CEO of Krugle, Menlo Park, CA, for his reaction to the big news today….that Google was launching “Google Code Search.”
Tech~Surf~Blog: What’s your take on this bomb, Steve?
Larsen/Krugle Inc.: We’ve been saying for 18 months that Code Search is an important new category. 
The massive benefits that can be realized by both developers and corporations when robust code and technical search is available is becoming increasingly obvious. Google’s entry, along with O’Reilly’s entry a few months back and the existence of koders.com — it all signals we’re on the right track.
Tech~Surf~Blog: What are some other reasons you believe that?
Larsen/Krugle Inc.: Well, for one, developers spend 20-25% of their time just looking for stuff. Krugle users report saving significant time due to us helping them find code and technical information fast. Second, code search is the key to making code-reuse a reality. Vast amounts have been spent to modularize and make code componentized, but little has been spent on how to find and connect the modules. Good code search fixes this. Thirdly, companies who care about developers — like IBM, Sun, Microsoft, and others — have spent the past several years attempting to provide better access to their code archives and high-value technical information. This is a big issue for them. We’re helping.
Tech~Surf~Blog: What do you have to say specifically about Google’s code-search product?
Larsen/Krugle Inc.: First, it could be characterized as a better version of koders.com. While it does simple code search, developers tend to be more exploratory in the search process. They first search and review potential answers, then research these answers to better understand the project and review the technical documentation. Our three channels — code, tech pages, and projects — are based on how developers really use search.
Second, searching and finding code alone is not all that compelling. To be truly useful to developers, you need to show code in the context of its project, the meta data around the project, as well as license information. Krugle does this with an easily browsable project directory and rich technical information that relates to the code and project.
Thirdly, search engines ultimately are judged on the accuracy and relevancy of their results. The same indexing and page-rank algorithms used to search text, images, video, and books come up short when faced with the demands of professional programmers. At Krugle, we’ve found the key to great search results is to parse the code, collect and analyze project information and statistics, and use all of this to more accurately order our code search results.
Fourthly — is that a word? — we think collaboration around code is very important. And we provide users with the ability to create, add, and save notes to code and sets of related results, providing a level of interaction with code that developers heretofore have just not had.
I even have a fifth point: While other code search engines offer the ability to search for a code file if you know its name or something about it, Krugle goes well beyond that by supporting common programmer use cases that revolve around: (a) exploring a new technical project or concept in an effort to quickly “come-up-to-speed”…(b) finding code or code snippets to use or to have as an example…or (c) helping find what’s wrong when your code doesn’t work and you need to know how to fix it.
There you have it, friends. Krugle should be liking all this attention. I think this code-search game is just now beginning to get interesting…
Tags: Krugle, Google Code Search, Koders, O’Reilly
Recent Comments