One of the turning points in history was when manufacturing embraced intermediate production. By creating simple components that can be integrated into complex products, manufacturers are able to build faster and cheaper, achieving better quality. In this tale of developer meets engineer,I describe how I'm using Python's inheritance model to bring this manufacturing reality to life in source code.
Programmers like vivid and clear statements. They work well as rules of thumb which can be used as justification of hard choices (especially bad ones but that's a whole different story!). One of those nuggets is: "Size is code's worst enemy". Very true. Unfortunately expressing any system of reasonable complexity requires vast amounts of code. What can be done to mitigate the risk of landing a codebase that is impossible to move forward?
A possible answer is: by using components, e.g. by dividing chunks of code into bits that can be freely composed together. In this talk, I discuss how multiple inheritance can be used as a composition tool which helps isolate unique bits of functionality. This leads to another useful principle of software developement: DRY (Don't Repeat Yourself). Done right, it can also be a way of creating truly reusable components.
The primary framework for real world examples will be Django, namely its ORM, forms and class-based views. However, the principles described are generic. Throughout the talk, I will mention details of implementing composable abstract models, form and view mixins. I'll explain how this approach creates readable and maintainable idioms along with pros and cons. I'll conclude by summarizing my experience trying to create a library of strictly reusable components.
lck.djangolibrary, e.g. abstract models handling time tracking, editor tracking, soft deletions, display counts, etc. etc. and how they can be freely composed together.