Archive for November, 2007


November 26, 2007

You’re invited!  DapperCamp is a free two day conference for developers, partners, Dapper users, and anyone else interested in Dapper.  The conference will take place on February 4 and 5 in downtown San Francisco.  It will include technical tutorials, hack sessions, competitions, prizes, feedback sessions with the Dapper team, and more.  Mitch Kapor will be delivering the keynote address and our partners will be presenting the ways in which they use Dapper (if you’re interested in presenting, please let us know and we’ll get you a slot to present).

Visit the DapperCamp website at The website is still under development – you can expect a collaborative workspace to be put up in the coming weeks where you can share Dapper ideas with the team and with fellow attendees.

Attendance is free, but spots are limited so we ask that you register for the event (registration is available at the DapperCamp website).

We’re still fine tuning the schedule, so it may change in the coming weeks.  In the mean time, your input is highly valued so if there is something you’d like to see covered, don’t hesitate to let us know.

Dapper Got Some New Magic

November 21, 2007

Hi guys,

Usually at Dapper, we don’t do a big song and dance every time we introduce a new feature or fix a bug, if only because we introduce features and bug fixes daily. Today, however, we’re introducing a big release – MagicBox – of which we’re especially proud.

MagicBox is first and foremost about your Dapper experience. Previously, in Dapper, there was a limit to the ways in which you could select the content you wanted to be a part of your Dapp. We’ve worked really hard to make that process much more flexible and scalable, and we hope, even entertaining. MagicBox is a new learning engine that infers from your clicks what you actually want to define as a field, and helps you converge towards your ideal definition rapidly. In fact, MagicBox’s learning engine is the first learning engine we’re aware of to be implemented in Javascript on a real product, and we’re very proud of it.

One of MagicBox’s important features is that it also listens to what you don’t want. If it highlighted too much, you can click on the extraneous items and MagicBox will suggest a new configuration that better suits your needs.

Our main goal with MagicBox was to simplify the experience, and using our new learning engine and the ability to “deselect” elements we’ve managed to get rid of the more cumbersome tools that we had before, such as the “sensitivity slider” and the “isolate” option. Now, to use Dapper, the only thing you need to be acquainted with is your mouse’s left button. Just give it a try and click away. If MagicBox didn’t highlight enough, click on the missing stuff and you’ll see how rapidly MagicBox corrects itself. If it highlighted too much, just click on the unwanted stuff and MagicBox will get rid of it in an instant. Don’t worry – just play with it – we promise that eventually MagicBox will always get you what you want (we have a strict money back guarantee policy :).

The second major update in MagicBox concerns Dapper’s backend engine. It just got faster, in some cases much faster (up to 10-15 times), more robust and more scalable. In particular our new backend will allow us to introduce new algorithms and abilities to Dapper much faster, and you’ll start experiencing it very soon.

Here is a list of the major updates in MagicBox:

– Clicking interface overhauled. Clicking page elements behavior completely changed, slider eradicated, new adaptive learning algorithm implemented. Result is that when you click something and it grabs too much, you just click the extraneous instances, if it removes something you want, you just click it. The algorithm learns with each click and approximates accurately what you’re looking for. This increases usability and robustness.

– Backend algorithmic infrastructure overhauled. This will make it easier for us to release algorithmic updates in the future. On some sites, we’re also seeing a 15x improvement in speed in Dapp execution.

– New algorithms implemented improving ability to extract content reliably.

– Bugs with back links improved.

– Borders in Dapp Factory switched to inner rather than outer, eliminating the case where you can click on the border accidentally.

– Timing issues with gray overlays, particularly when editing a Dapp, resolved. Fixes the bug where you’d edit a Dapp and it would have a Javascript error, requiring a reload of the page several times.

Some known issues remain and we’re working hard to resolve them. We can’t do it without you, so please let us know when you encounter an issue.

Happy Dapping!

Interview about Dapper on BungeeLabs

November 17, 2007

A few weeks ago Ted Haeger and Alex Barnett at BungeeLabs interviewed me about Dapper, our technology, the Semantic Web, content ownership, and business models. The interview was posted today to their blog and if you’re interested in hearing more about what we do, check it out. Check out the interview here. Thanks to Ted and Alex for a fun conversation. Don’t forget to check out their other episodes – they speak with great people in the web-services area (Jeff Barr from Amazon, Jeremy Zawodny from Yahoo, and more).