They were even able to gather a few prizes to raffle off. The more contributions you made, the higher the chance to win something. About 120 tickets were solved during this joint effort and a few new bugs were discovered, so I’d say that this initiative was pretty successful.

I was able to commit 6 patches to both 2.3-stable and master branches. A part of me will be in Rails 3.0! Neat, huh? I had lots of fun digging around mostly through activerecord and actionpack. I also had the opportunity to get in touch with git — branching, patching and whatnot. Lets see if I can win some cute books!

Dear Hugo Peixoto,
We have the pleasure to inform you that your contribution entitled
NeatSqueak on wheels: Neural networks applied to movement optimization of a simulated robot (paper #131)
submitted to EPIA’2009 has been accepted for presentation at the conference in Aveiro, October 12-15, 2009, and publication in the conference proceedings.

Somehow, we’re in. Our first NEATSqueak related paper got accepted!

This paper is the result of our work done during the Robotics course. It focuses on the application of unsupervised learning methods to the Ciber-Rato competition. We developed a neural network to control a simple robot so that he follows a set of points.

I don’t know how it got accepted; we were severely bashed on most of the four reviews and our work was very simple, not even suitable for a conference. We were just aiming at getting a good grade on our course.

Now we need to review a couple of issues, and we’ll be presenting it in October, at EPIA’2009. I really hope we don’t get flamed much during the discussion.

This mechanism, although promoting usability,  leads to the usual trust issue: you have to give away your credentials to a third party in order to use the service.

Now, although we’re pretty much honest people regarding these kind of details, one can imagine the damage that could’ve been done if we were storing the users’ credentials. Several teachers are using our service. With this in mind, we had to think of an alternative approach.

Here enters OpenID:

OpenID is a free and easy way to use a single digital identity across the Internet.

This is exactly what we were looking for. Basically, the website that requires authentication redirects the user to his OpenID page, where he can confirm his identity with his credentials. After being validated, the user is redirected again to the first website, carrying a token that guarantees his identity. This way, there’s no need to enter our credentials in the website requiring authentication.

OpenID also has some extensions (SREG and AX), which allow the identity provider to automatically fill in registration fields, such as email address, full name, country, etc. Of course, this requires the user’s permission. Or at least, it should.

So, me and Pedro Coelho proposed implementing this service so that it can be used with our faculty’s credentials. Obviously, it will be deployed and validated by our IT department so that the users can trust the service.

Yesterday, we completed the first stage — we achieved OpenID 1.1 full compliance:

OpenID@FEUP current status

Now, we need to continue and implement the 2.0 features, along with the SREG/AX extensions. We’ll also need to work on the user interface a bit, so that it doesn’t look like this:

OpenID@FEUP user interface

In this case, OpenID can be used not only a centralized identity, but also as a way for our services to verify if a user belongs to our institution without asking for their passwords. I’m hoping that we have this finished by the end of the semester, so that it can be integrated with feupload and other applications before the next school year.

With this in mind, I’m planning on writing small and simple posts on what I’m currently working on. An OpenID provider implementation at FEUP, an XML dialect for UML class diagrams with an SVG display, A.I. experiments, or something else that randomly pops into my interest.