Who You Calling A Jesse?

Forget the Starbucks popping up all over and the yuppiefied uptown core of the city, if you are looking for signs that Waterloo has a potential to really take off in the world of innovate web application development aideRSS is one of them. This site is just cool and relatively simple in its concept from what I can tell. It takes all your RSS feeds, ranks them based on the community reaction to the post, and gives you an optimized feed option so you only see what is worth reading in the eyes of the larger community. It seems to be a little like Technorati (or maybe a lot) but focused on your specific RSS feeds not just your blog.

The Techcrunch community seems to have a lot of positive things to say about it as well. I really hope these guys can make it out to the next BarCampWaterloo.

It has been over a year since I made a decision (inspired by a student, Mitch) to switch development focus from PHP to Ruby on Rails. It all started with the need to build an events management system in a way that other co-op students can pick up and understand quickly. We now have two version of UW Events, Kiwi, Opinion, Experts guide, a conference registration system, and a new app based on Twitter that is known as chatter (still very much a development version).

First the advantages; It is dead simple to manage, gems are just amazingly helpful, the community is great, applications are portable, stable, and expandable. All things you probably have heard from one source or another. What I find as the biggest advantage is that the Rails framework ideal for an environment where you have a different co-op student coding every four months.

What I mean by that is the framework enforces consistency in style and encourages code reuse. These two things have been the biggest problems when you have a new developer come in that just doesn’t code the same quality or style as the previous developer. Some are learning as they go (CS students that can code Java not Ruby), others are experienced and have their own style. That is not to say a PHP or ASP or Python framework wouldn’t achieve the same thing.

Related to the framework advantage is that it enables me, the manager, to be able to relatively easily jump into the code and check the quality. I haven’t done that as much as I should but that is because I haven’t seen ugly Ruby code yet (in just over a year) so I don’t feel the need to spend a load of time checking up on the student.

Now the negatives; server end support on campus is not ideal (not officially supported), other campus web folks aren’t using it, Dreamweaver and others tools of choice for front end don’t work with it, students aren’t confident with it, and the gems can let you down at times. The last point is actually the worst one for one big reason—you will hear ‘oh its easy just to rewrite it’ from the developer.

The moment you hear that is the moment you know the framework ideal is weak with this developer. You need to get back on top of things and see why the developer/student is thinking it needs a rewrite. It is a time sink to start mucking with gems so don’t do it unless it is absolutely necessary.

The server support probably needs an explanation. Server wise it is one part rails one part learning curve. Rails is easy if you just stick to fcgi, relatively simple sites, and lower traffic. The moment your traffic spikes is the moment you need to think about this differently. With our latest project we have taken on the whole mongrel cluster world, dove into proper subversion management, and use capistrano to publish the application. That learning curve is rough but worth it. It has certainly reduced the weight of the negative.

Fourteen months later I am still in love with Ruby on Rails even after the romance starts to fade.

We have been sputtering along a bit recently between vacations, newly acquired meetings, and other things here and here but UW Chatter (twitter clone) is still moving along with an eye for a stable ‘production’ launch in September. Starting soon UW’s Housing and Residences will make an effort to push out content via Chatter and we will see how it goes. So what is left besides bug fixes? Well an API.

I would like to see Chatter act as a communication hub for web applications on campus in much the same way Kiwi can be an authentication hub. A simple messaging API is easy enough but how one that:

• handles what group you want to post to remotely
• send direct private messages to one or more people (if their preferences allow)
• control your preferences for that particular application outside of Chatter and have that application honour your notification settings in Chatter

It starts to get complicated but Bobby and I have had some pretty good conversations about it. The issues of what to use (XML-RPC, SOAP, etc) with an eye fixated on REST is making this a little more focused but really it’s all on the table. How do we manage this?

Well we do have an idea ;) That will be another post next week… for now I am curious what people think need to be in an API for any web application never mind a twitter clone. A note on comments, I am on the road so approving them might be slow.

Today I finally uploaded an updated version of UW’s International website. It was a long time coming and its design is a collaboration between Matt Regehr in Graphics and myself. Matt did the heavy lifting design wise, I just offered some CSS skills and general design direction. Kevin Paxman from Graphics also offered some help.

The layout is very close to the original design for UW’s Common Look and Feel—before it was stripped down to bare bones and had a template scheme to make it easy applied to it. With it being wider (900px vs 680px) we took the opportunity to place the navigation further up the page and highlight the sub navigation better. The div class naming isn’t following convention (yet) but it was a bit of a rush between vacation schedules, content being agreed to, and the design coming to life.

I think it is a big improvement over the previous site that was based on the much older standard I developed in 2001. Hope people like it ;)

The list of web apps for the iPhone keeps growing with some mixed results reported. 37Signals went and announced their customized for the iPhone tada list, which is great but with all this it has me thinking—what about the other 99% or so of mobile users out there? All they really need is a solid browser that supports javascript like the iPhone does I would think. Could Opera Mini do that? I certainly hope so. Take a look at the iPhone interface in JavaScript to see what might be possible.