Something I haven't seen anyone here bring up yet: many URLs are meaningless to humans past the domain. For an example, look at the top of this page. The only semantic meaning in the URL for this post is `news.ycombinator.com`. The rest, `item?id=24156986`, is meaningless to a human. (But, of course, meaningful to HN's backend.) A lot of (most?) of the URLs on the web are not semantic. They're naked application look-ups. I claim that for this current page, there's no meaningful loss of information by just showing the domain.

But some URLs are semantic. I think losing those would lose useful information for human readers. If we could perfectly know which URLs have useful semantic information for users and which don't, and then only present those with full semantic meaning, I wouldn't mind much.

Main point: I see people saying things like "these designers think people are too stupid to understand URLS." But that ignores that some URLs are not actually meaningful to anyone.

Anticipated responses:

1. You are losing information by taking off the "application lookup" part: the information that an application lookup was made. Fair enough. But I claim it's a small loss.

2. We can never perfectly separate out the URLs with useful semantic information. Which, also probably true. But I think we can do a decent job, and as long as the full URL is present when I mouse-over it, I probably wouldn't object.