Fork me on GitHub
Subscribe to RSS Feed

Kit Plummer
Software Engineer :: Techitect :: Evangelist
kitplummer@gmail.com
kitplummer (AIM,Yahoo!IM,Gtalk,Skype)

To the top »

Please share:

The quagmire that is the world of software frameworks is growing on a (almost) daily basis. And each language has its own pile of options. Java, Python, Ruby, PHP (Smalltalk, Erlang, Perl, Haskell) and the like all seem to have a play. I don’t want to waste anyone’s time … this post is not a quantitative review, just some anecdotal commentary.

At the genesis of the rVooz project we were a little quick out the chute and started in a direction that was more familiar to our “enterprisey” skills. It didn’t take long to realize that we’d rushed in without properly considering all of the requirements (at least the one’s we knew).

Modularity desires led us in to the realm of OSGi – and we were really comfortable moving in that direction due to our familiarity with Java, Maven, and Eclipse. Unfortunately, a desire to architect around REST interfaces (implying HTTP services) conflicted with the modularity from a couple of fronts. First, thinking services when were building a server. Second, thinking modules when we really wanted scalability – both horizontally and vertically.

The architectural forces were shifting on us, and we needed to jump ship on our framework choice fast. Thanks to this gem we decided to follow suit and construct a matrix that led us, obviously, to the same conclusion. Ruby and Rails has been an interesting step, a smart move for sure. We’re also playing with <a Ramaze to get a feel for what else is available in the Ruby landscape. It is very nice too.

The things we’ve been able to accomplish in a very short time with a few developers are pretty amazing. However, it’s much more than just looking at a language, and a subsequent framework. Your architectural ambitions (Internet scale Vs. a few desktop application Vs. a personal blog) are very important. Your project specific requirements are equally important (i.e. our need for geographic data and processing support). Make sure you weight the most important aspects as appropriate – to help provide separation between like frameworks.

A couple more hints. I would also recommend that you place less of a burden on the language you are currently comfortable with. It is much better to think “the right tool for the right job” than busting out the ole hammer – even when there’s the potential of an immediate learning curve. Take a look at the community around both the framework and language. Don’t get caught up in the “hip” factor either – make sure you can really see the culture for what it is. If it is a commercial offering ensure that you understand the company (and its life expectancy). If it is an Open Source project spend some time in the trenches chatting in IRC. Try and find out the formality of the products/projects testing and release cycle and get a feel for the objectively and drive that exists. Lastly, don’t ever be afraid to admit a mistake or fear course correction. I could quite possibly mean impending doom.


Please comment:
blog comments powered by Disqus