an ordinary blog — andr3.net

A week ago today I was arriving from a weekend spent amongst the Portuguese geekdom. I attended my very first Barcamp, in Coimbra. So, in the spirit of the event I prepared a talk about something that have been on my mind recently.

Instead of repeating the presentation about microformats, which is nothing more than an attempt to push adoption of microformats to increase semantics on websites, I decided to look at it from a different angle and show how you can muster the value of semantic content, available TODAY on our users' other websites/services.

If you're interested, check the presentation below. I gave the talk in Portuguese, but since there were so few slides, I translated them to English.

Feel free to give some feedback in the comments below.

Demo and Code

For those of you who didn't attend, I wrote a tiny little script over two nights that acts as proof of concept. It simply scrapes a given page and tries to gather data of the user's attention profile.

The goal would be to be able to provide content in which the user is interested right off the bat, as soon as they sign up. I described the harvesting algorithm in one of the slides, but here is a small list of the steps:

1. Get page contents.
2. Look for an APML
3. Gather set of rel-tags
4. Scrape any rel-next, rel-previous or rel-archives.
   This step is not done in the script, it just detects the presence of the links.

You can test it for yourself at: http://workshop.andr3.net/tageater/

If you want the code, it's available for download as zip file. I didn't put it up on SVN or anything because it's just an example. It's written in PHP (5.0), requires curl, DOM and SimpleXML.

It uses the microformats transformer Optimus, written by Dmitry Baranovskiy.

It includes a very basic cache mechanism, based on the filesystem. Make sure the ./cache/cache folder has the right permissions, if you want to use it.

I'm releasing all this under the MIT License, but remember this is only a proof of concept, or in another words, do not use this in production.

Examples

Here's a few tests I made with it. First, and this is included in the script eve though it's deactivated, I pointed it at the all the URLs specified by the BarcampPT attendees. On top of that, to avoid skewing the results by grabbing URLs owned by fellow geeks, I pointed it towards 5 of the highlighted blogs at SAPO Blogs at the time.

Q&A

A very important question came up from the audience right at the end.

What's the motivation to open up and publish our users' attention profiles?

By opening up your platform and allowing the users to knowingly publish their attention profiles, you are not only becoming a part of the web's ecosystem. You are increasing the value and importance of the users profiles in your own service/website!

Or as I put it, when the user signs up on any other website or service and sees that it's possible to import his attention profile from YOUR service, who do you think will benefit? Both the new service AND your own! He will become aware of the increased usefulness of having a profile with you. More here.

If you have some more questions, shoot. :)