Why I need to develop a new Python Web framework

I begin developing Web applications on 2004 for the library at UCLV university, back at that time I use PHP as programming language. But 5 years latter it became clear for me that the same design decisions that make PHP a easy to use programming language for Web development purposes where the same that degrade exponentially the performance of my products as they became more complex.

On 2009 I met Python and PEP 333 for the first time and the first impression was: “Oh wow! This is what I was looking for!”. The simple concept of application presented at PEP 333 and the realization of its capabilities make me switch to Python as programming language for my future Web experiments.

After toying with most of the open source Python based Web frameworks I realize of some issues that arise on them:

• The use of thread local (global in the worst cases) objects for sharing resources/services. This issue make the testing of components complicated.
• The use of the Web request object as a registry for sharing resources/services. This create a dependency problem and corrupt the meaning of Web request.
• The use of the WSGI protocol for connecting required application components as middleware.
• The use of the WSGI protocol for writing actual Web request handlers. This force the developer to handle the complexities of this low level protocol.

The first two issues share a common consequence: component requirements gets replaced by the framework (because you need the framework to provide the component requirements trough the use of a global object or the request object) and the real component requirement expression is at the component implementation and not at the component interface. This break the rule that explicit is better than implicit.

This are the conditions that motivate me to develop a new Python based Web framework with PEP 8 as coding style and the Zen of Python as design philosophy.