Motivation

• works disconnected from backend (i.e. the repository server):
• load caused by public readers doesn't influence performance of repository for authors
• minimize potential security problems
• high performance: all (most) publishing work is done beforehand
• independent from user: easy load balancing across multiple servers and easy fail over (no service interruption)
• allows deployment on widely available hosting infrastructure
• easy, understandable architecture
• still some dynamism possible (server side includes of static or dynamic content, redirects configured in web server, <your own thing>, ...).

[more coming]