I’ve clearly been neglecting my little spot on the web.
It has gotten so bad that Brendan Eich had to link to a web archive copy of a page of mine. I must say, however, that it is very ironic and amusing that it is was that particular page. The problem turned out not to be a software problem, but rather a (presumably inadvertent) DOS attack on feedvalidator.org, causing CGI processes to fail. Blocking the IP address in question caused the problem to clear up.
General outline of my current approach:
- My interface to my weblog will no longer be Python/CGI application on a hosted server. Instead it will be a Ruby/Sinatra application on my private home server where keeping things up to date is much easier for me. That application will produce static HTML, CSS, StyleSheet, and a single feed, all of which will be rsync'ed to the public server.
- The only services exposed will be search and comments. Comments initially be disabled, and when they return they will likely be moderated, though I may make the moderation queue publicly visible.
- My current focus is a software update. The overall look and feel will (at least initially) remain the same.
- The pages produced will be HTML5, though all pages may not always pass validation. Mike is 100% correct: different people can make different judgment calls. In particular, I continue to find that explicitly quoting all attributes and explicitly closing all elements both reduces authoring errors and enables a wider variety of user agents to parse the pages correctly.
- I’ll likely drop many features that were popular at one time, but no longer appear to be. An example of this: OpenID.