Migrate LLTHW Web App to Asynchronous Server
The user experience for LLTHW could be vastly improved with an asynchronous web application; currently the application runs on Hunchentoot—a thread-per-request server—which is inherently greedy in terms of system resources. While useful for web applications that need to validate and process arbitrary user data in an isolated environment, LLTHW only needs to serve up content. An asynchronous web application will thus be able to maximize available system resources and provide a faster, stabler experience for all users even under the heaviest traffic spikes (such as the Hacker News fiasco this past July), and buffer against DDoS attacks.
This will also allow for a more sensible use of 3bmd to provide live-updated content, without specifically re-parsing each Markdown document for each user on request; only one copy of each parsed Markdown document needs to live in memory, which can be accessed sanely by all clients. As such this ties in with issue #3 (closed) to reduce application memory usage.
The most mature Common Lisp asynchronous server and web application framework appears to be Wookie. It is still in Beta, but it seems to be worthy of production use-cases, such as the official Wookie documentation, and the Turtl platform.
Alternative suggestions are welcome.