Matt Gemmell

My new book CHANGER is out now!

An action-thriller novel — book 1 in the KESTREL series.

★★★★★ — Amazon

Burying the URL

tech & interface 3 min read

Last week, one of the nightly builds of the Google Chrome web browser experimented with hiding the URL in the address bar, instead showing what Google calls an “origin chip” - essentially a button for the domain you’re currently on.

It is just an experiment, and it has some potential usability problems, but it got me thinking about our old friend the URL.

As an implementation detail, URLs are probably never going away. We clearly need an addressing and routing system in order for the web to work1. Domains provide disambiguation, brand identity, and are already embedded in popular consciousness. They certainly work.

URLs aren’t without problems, though (and here I’m just talking about web URLs, for simplicity). There are plenty of things that are confusing about them:

  1. Most people have no idea how they work, or what they mean, or how error-resistant they are (which is to say, not at all).

  2. Most people don’t type URLs, nor look at them whilst browsing. I expect that most people don’t even use bookmarks, and instead simply use search engines to reach their commonly-used sites.

  3. They’re ugly and artificial, exposing part of the routing architecture of the internet with no concrete user benefit.

  4. It’s somewhat confusing that they progress from the most specific to least specific information in the main part of an URL (www. apple .com), but then become increasingly more specific after the “/” character. There’s an abrupt change of direction. Sorted more sensibly, in order of increasing specificity, URLs would look more like “com.apple.www/directory/filename.html”, and indeed that’s a very rough reflection of the lookup process. But they’re displayed (and entered) in a haphazard, muddled format instead.

  5. The vestigial and awkward “www” still hanging around in some cases, for no clear reason. Similarly, the differing (and essentially arbitrary) global top-level domain suffixes like .com, .net etc are of nebulous origin and purpose to almost everyone.

  6. The scarcity of desirable traditionally-suffixed URLs has lead to an explosion of hyper-inflated new suffixes, which is a step along the road towards completely arbitrary and free-form domains.

It seems to me that search has long since become the default navigational technology of the web, with the main user-focused enhancements coming from predictive and historical results, automatic collections of commonly-used sites, and so forth. By contrast, hyperlink navigation and explicit curation of bookmarks are for the tiny minority. Humans just don’t use the web that way.

Search has issues too, of course. Disambiguation is the main one conceptually, though it’s rarely a significant problem in practice (particularly when personal navigation history and information can be leveraged). Impersonation is another, which is a special case of disambiguation. The user-experience friction of having to choose what result you want is worth mentioning, though again I don’t think it’s much of a hurdle in practice.

We’ll always need URLs behind the scenes, but it’d probably be good if they took a step back out of sight. Safari, Chrome and Firefox already visually soften all but the domain portion of URLs in their address bars. Advertisements increasingly urge users to “search online for” specific terms or names, rather than to visit a given domain. URLs have always been difficult to enter accurately, and never more so than on touch-screen mobile devices.

From the visitor’s point of view, this article you’re reading is something like “Burying the URL Matt Gemmell”, or almost any meaningfully specific subset or rearrangement of that phrase. The title (on page; not necessarily the HTML title in the markup) is how we conceptualise the content. The address is just some technical trivia, that seems somewhat at odds with the modern internet.

URLs can and do change, and as long the change is performed gracefully, nothing is lost. The content is the thing, and the convenience of the user has to be the goal in mind.

Perhaps it’s time to acknowledge that the web’s infrastructural user experience should emphasise the humans who are doing the browsing, rather than the machines serving up the content.


If you’re interested in discussing this further, you can find me on Twitter.

My writing is supported by readers like you. Any contribution helps enormously.

  1. Though hierarchical addressing is really just a technological stalling tactic, not an optimal solution - and certainly not from the user’s point of view.