So there's artists that just link to some social media website, not considering that accounts can be suspended, usernames can change, etc.
Similarly there's also developers that "link to"[1] dependencies without considering that repositories might disappear (together with the source code for that dependency's version if nobody backed it up), a package's version might be removed from registries, online documentation for a dependency could disappear (ugh), etc.
[1]: Just adding name+version to whatever manifest file and forgetting about it forever. Maybe adding a cache (not even a proper mirror, much less any self-sufficient way to build the dependency in case of disaster).
This doesn’t always work because many websites neglect page titles. I’ve always wanted to ask a wide range of web developers why they neglect titles. Why?
The fact that people use the HTML title attribute for everything except for including the proper title of the thing pointed to is sort of perturbing. Even for Google SERPs and here on HN, it would be useful to have access to the full thing instead of truncated titles when they appear, but neither site is an exception to the tradition of non-use.
Your idea seems fun if people can one by one describe what the page use to look like and/or crawl the web for clues. We can have our under reconstruction banners back.
Sure it would eat some bandwidth, but so do prefetch schemes and other stuff intended to speed up page renders.