I’ve learned many lessons from recovering weblog posts from the Wayback Machine.
If you feel the need to freshen your site, you really don’t need a new domain or subdomain, and most of the time you don’t need a new site redesign. All you need, is a break.
It’s OK to mix content types. So what if your techie stuff is side by side with your poetry? People can skip what they don’t like.
Don’t ever embed a dependency on a third-party web site for anything. So many of my recovered posts referenced photos in an account on Flickr that I no longer have. I recovered most images, but not all.
If you’re going to embed a tweet, embed a screenshot of the tweet and then provide a direct link to it. That way if the account is gone, you don’t end up with gibberish in your page.
For your own media, be wary of using the weblogging software add media functionality. In the process of recovering my writings, I decided to go from a WordPress multisite to a single site installation. It wasn’t a complicated task—export site’s content, install a fresh WordPress installation, and import the site—but the export/import munged the media I had added using the WP Add Media functionality.
(Thankfully, there’s a plugin, Broken Link Checker, that helped me find the broken image links and get them fixed.)
Don’t use excerpts on the front page, or the category page, or archive. I lost some writing because the Wayback Machine had a post in the main page, but hadn’t captured the actual post page, itself. When I displayed the full contents on the main page, I was able to recover the writing. When I didn’t…well, that writing was lost.
Lastly, even when recovering old posts I still left internal links to the Wayback Machine. So much of the older stuff is gone but still preserved in the Wayback Machine.
And if the post had comments, I included a link to the Wayback Machine entry at the top of my posts so that folks can see it, and the discussion, in its original form.