Archive for November 26th, 2008

Big Update to Dapp Factory

November 26, 2008

We’re pleased to announce that we’ve released a major update to the Dapp Factory.  In the new version, you can now see how the content selection works across multiple pages and fix it if it isn’t perfect.

For some background: Dapper works by providing it sample pages that have the same structure, but different content.  For example, if you want to create a Dapp that allows you to extract search results from your favorite search engine, you would provide several different examples of searches (e.g., “cat,” “dog,” and “meow meow”).  Dapper will use the different sample pages to understand the structure of the page, so that later you can use it to conduct any search you want, even one you didn’t train it on (e.g., “woof woof”).  In the past, our interface allowed you provide multiple sample pages, but when it came to selecting content, you could only see how it worked on the first sample page.  If it didn’t work well on other pages, you would only know that after you saved your Dapp, and editing the Dapp so that it worked better on other pages was a chore.  With the new version, you can now see when you are creating your Dapp how it works on all the sample pages provided.

To help illustrate how the feature works, here’s a screencast demonstrating it:

We hope you find this new feature useful and that it helps you create better, more robust Dapps that work across many pages. Should you have any questions or want to discuss it with us, don’t hesitate to contact us.

Something to note: the new version does not extract content from all the pages at once – Dapper remains true to the concept that it extracts content from one page at a time.  Combining content from multiple pages is something that is left to techniques outside the Dapp itself (e.g., MS Popfly with their awesome Dapper integration, Yahoo! Pipes, or good old fashioned programming).

Advertisements