Serverless-wiki is a 'serverless' public wiki
This means most of the functionality can be accessed while serving static pages, and only for updates is any custom server-side code required.
In a nutshell:
See http://serverless-wiki.s3-website-eu-west-1.amazonaws.com/
'logging in' is initially a usability feature rather than a security feature: the password is put though a non-cryptographic, high-collision hash function.
The 'real' authentication happens when POSTing new markdown.
Design space:
To be fair an attacker could still first brute-force the private nonce and then do a dictionary attack after all, but especially for a public wiki this seems sufficient.
The actual HTML generation takes wiki templates and markdown sources as input, and generates HTML output.
We expect a low throughput (after all we hope to be mostly serverless, so also keep Lambda invocations to a minimum), so we'd like a platform that has fast cold start times. While the JDK can be a fine choice for high-throughput Lambda functions, it might not be great for us. Nodejs and python are comparable, and given that choice I prefer python.