Unbuilt systems

In reading about architecture (buildings, not software), I realized that for architects and the history of architecture, the “unbuilt” buildings play a significant role. I guess that the most common procedure when offering on a project is to let architect firms present their architecture, along with cost estimates, etc, after which the most appealing solution is chosen.

This probably occurs in software, but it can’t be that common. And if any architectural work is done when offering, it certainly isn’t included in the proposal, since the buyer of a system most likely isn’t able to weigh different architectures against each other and determine which is the most suitable.

But the idea to construct systems that aren’t meant to be built – as a learning experience – really appeals to me. It could either be a completely unprecedented system, or an alternative to a system that one has participated in building, and for which one might have lots of ideas about what could have been done better.

The other day I found Great Buildings Online – an archive with lots of buildings, with photos and free 3D plans. Imagine such a database for software systems – both built and unbuilt – with related discussion forums. Programmers could collaborate to find larger-scale architectural patterns and document these.