[superkuh]

It is a bit atypical because I wanted a purely static comment system for my static site. Using nginx to just log certain matching URL types to a custom log file was way safer and easier to maintain than anything written in a dynamic language accepting arbitrary internet input. I eventually adapted this approach to receiving webmentions as well and in that context I provide textarea input for webmention response URLs.

My method is not better against spam bots but it does suffer from a different type of them. For example, if a spider follows a link it sees of an existing comment, "...com/something.jpg/@say/..." then my system will interpret that as a comment. So... spam bots don't see it, but normal spiders can cause spam.

[rambambram]

Thanks for explaining. I'm in the process of collecting interesting websites, either as a bookmark, but better yet, as an RSS feed. I can't seem to find any feed on your website!?