Choosing a web framework

Read my About page to see what I’m up to.

I’ve been using the major frameworks for Java professionally.

It started with Struts. Then came WebWork. I did something small with Spring Web Flow until I went to a company that made the decision to jump to JSF.

It has been an interesting ride. The biggest annoyances with web development have been:

* It takes too much time. Sometimes a small thing takes days to solve. Most time goes to trial and error weird working tags, xml configurations etc. The server side development is nothing compared to these issues. Searching the web for datatable and commandbutton will give you an example of an issue that has made many guys bald. The development cycle is also slow. You have to restart the web server when you change a JSF configuration. When the application grows, it is frustrating to wait for the server to restart after every minor change.

* You need very experienced developers. .NET-developers in our company seems to get more done even without years of experience. I think this article is partly right.

So I decided to choose Turbogears as the web framework. First because I want to learn Python. Second because Turbogears is a collection of pre-existing projects. I believe these projects must be architecturally in a good shape, since they can be used as modules in Turbogears.

There’s also a lot of fuzz about Turbogears. Not as much as of RoR, but enough to get you interested. The biggest claim is that it is fast to get things done with. Sounds good, but I want proof.

I found some interesting links when researching Turbogears. Bruce Eckel is also trying out Turbogears. Eckel is the author of a book that thought me Java in a great way.