The Internet Archive Wayback Machine is great for scraping content and saving it, and also for serving said content through a new archive.org URL the latter part of which reproduces the latter part of the source URL.
Proposed is a high-level deal with the domain registers and authorities such that an abandoned website can be scraped and also its URL redirected. I'm thinking that a similar user interface might be used as a t present with a timeline floating at the top of the page showing previous captures of that URL. But the original domain would also be "captured" and so the original link from 15 years ago would still work.
What about domains which continue in use but are radically changed in their content? Well in theory the same thing could happen, when you enter an obsolete URL instead of getting a 404 page (or worse a silent redirect to the home page) you get hopped to the archive scrape and timeline while still being at the original URL.
What about individual URLs which are still active but have changed radically? Well in that case it would be up to the site owner whether to allow the Archive to put a floating timeline along the top.
Perhaps some at least of this functionality could be done client-side. But I would still like to see obsolete domains captured and vested in the Archive so they can be reunited with their scraped content.-- pocmloc, Oct 11 2020 https://www.cloudfl.../5xx-error-landing/ [pocmloc, Mar 22 2021] Further discussion dead link auto-resurrection [pashute, Dec 25 2023] So you’re proposing that when you try and open www.oldURL.com, instead of going to a ‘page not found’, it automatically goes to www.archive.org/oldURL.com? How does this mechanism know that the original request has failed?-- hippo, Oct 11 2020 The name servers point oldurl.com to the archive server-- pocmloc, Oct 11 2020 Came a cross a kind of half-baked implementation of this, I clicked on a link to visit a website, and I got a Cloudflare landing page which told me that the site was offline, and after a short pause it served me an archived version of the page with a header banner explaining this.-- pocmloc, Mar 22 2021 If you could map base64 onto UTF-8, and were allowed to encode your URL in an arbitrarily long UTF-8 string, you could encode the contents of your entire web-page as a base64 representation embedded within the URL. It might lead to less than optimal human recognition for a given URL, but it would make concrete the link between index and underlying content. Since any conceivable web-page can be encoded as a base64 number, you might not actually need to host your website anymore, since the whole thing would be reproducible from its (admittedly long) address.-- zen_tom, Mar 22 2021 Great idea [z] but not quite it
I remember [Vernon] posting images here in the link thingy using this method.-- pocmloc, Mar 22 2021 Yes possibly.-- pocmloc, Mar 23 2021 random, halfbakery