The Decline and Fall of The URL

The URL is a very powerful concept; it represents a universal way to access any resource anywhere in the world. Here’s one of them, as it appears in Firefox 5’s address bar:

Address bar containing

The first few letters before the colon are called the protocol, which tells the computer how to interpret the rest of the URL. The http protocol is the most common and specifies a resource on the World Wide Web, while the tel protocol specifies a telephone number, and https specifies a resource on the Web transferred over a secure channel that can’t be eavesdropped. Those are just a few; there’s lots of other ones.

Many user interface designers for browsers believe that most users don’t understand what a protocol is, which is probably accurate. Google Chrome’s solution is to hide the protocol when it’s http, but to display the protocol in all other cases. Firefox is now adopting the same behavior.

There’s a number of things that trouble me about this approach. I’ve already written about the behavioral impacts, whereby user expectations of copy-paste are broken and a confusing mode is introduced. Furthermore, because protocol information is still displayed for any non-http resource, understanding how to read the address bar is (ironically) made more complex.

Aside from those concerns, however, there’s something else I’m worried about. The main argument I’ve heard against exposing protocol information to end-users is that, if we present it, we might as well present all kinds of other information about the TCP connection, CPU registers, and other obscure technical statistics.

Now, I know I’m biased because I’m on the Hackasaurus team and trying to teach people the basics of HTML and CSS, but browsers have historically been very friendly to learning web-making, in part because they keep protocol information in the address bar. My guess is that removing the http:// neither helps nor hinders someone from using the basics of the web—but it definitely makes it harder to learn what hypertext is.

Understanding technology is relative. Someone can know the basics of writing an anchor tag without knowing what TCP/IP is, and it’s still quite empowering, in much the same way that it’s empowering to know how to grill vegetables without necessarily knowing everything about the chemical reactions taking place underneath.

Doing little things in the interface that promote transparency and help people move from being a web-user to a web-maker is important, so long as we don’t make things difficult for the people that just want to be users. I’ve never found my parents or other non-technical users to be confused by the presence of http://, which is part of why I don’t see much gain in removing it—especially given the behavioral shortcomings of this change. Far more exciting to me is the exact opposite approach: designing experiences to help users understand what the URL in their address bar means, and encouraging them to create things on the Web instead of just browse.

22 Replies to “The Decline and Fall of The URL”

  1. I dislike this Chrome clone too, fortunatelly this time it exist a preference in about:config to restore the protocol
    browser.urlbar.trimURLs -> false
    But I agree with you, it’s better if protocol is visible by default.

  2. I think I’ve been bitten by the copy/paste behavior about five times in the last week, mostly when posting a link on hacker news. (It’ll auto-format if the protocol is present, but won’t otherwise.)

    Reading this it occurred to me to look in about:config. I believe “browser.urlbar.trimURLs” is the setting to adjust — at any rate, I now have my “http”s back. 🙂

  3. I’m also concerned about the direction we take with urls. On top of what you said, URLs give additional value in bringing a reward for learning them – they give control. Yes, you have to learn URL’s, no not many people will, but those who will, and the group is much broader than just geeks/IT ppl gain more control over how they use Internet.

    They suddenly can change the url to get to the main page of the site they’re viewing, they can copy&paste the URL to someone, they can see if they use https or http (or any other in the future).

    The direction of minimizing users ability to view and manipulate the urls is giving more control to the web authors who slowly have to worry less about users who do not follow the paths set by the author.

    I find the ability to stop, reload, bookmark, go back, manipulate URL and access the web page (to modify it via some platypus/greasemonkey script) the fundamental right of the user that gives him control over how he experiences the Internet. He may not use it, he may not need it now, but maintaining it secures the power the users hold. And I’m worried that the direction IE9/IE10 takes, and sometimes Chrome, is to minimize the interactions and manipulations the user may do to the flow and URL.

    Of course the goal is to protect the user from having to think about too much and from making mistakes that may break the website for him, but the result may be that center of control will shift back toward the app maintainers, not users 🙁

  4. Absolutely agree. Decentralised addressing is crucial to the spirit of the web and user empowerment. Others have incentives to blur the lines and encourage users to rely on them to intermediate and provide centralised search/lookup/addressing. Mozilla doesn’t, and I hope pushes back and provides a URL experience which users enjoy and understand.

  5. Let me get this straight–you’re titling your article “The Decline and Fall of the URL” and it’s abut whether or not a browser displays http: in the address bar? I don’t have the time here but basically every sentence in your post ranges between so-what and ridiculous. The display in the browser bar “helps people learn what hyptertext is”??? Seeing the http:// helps me learn to create rather than browse? What??? Did you forget that even Tim BL regrets the double slash? Can you concentrate on actually making browsing and creating web content easier instead of on arcana?

  6. The problem with the address bar is, that many users (the most I assume) simply ignore it. Indeed they won’t even miss it, when it vanishes totally.

    I once talked to the IT guy of a customer about a specific HTTPS problem on their website. I couldn’t find out, what his problem was, until I finally figured, that what he meant with “standard way to access the page” was entering his company’s name in the Google search field on his home page, which lead to the HTTP version. He was, mind you, the IT responsible.

    To put it in a nutshell, it is a long, hard and in the end fruitless way to teach the people things they are not interested in learning.

    For cars we have the driver’s license to guarantee a minimum of knowledge before gripping the wheel. So if we do not force people to learn about URLs, what are incentives to make the “Joe Average” *want* knowing? Some people might start being interested after a phisher emptied their bank account, but there must be better ways. However, I cannot see them at the moment.

    PS: @Bob Myers: Calm down. Start coding HTML for some years and you see, that Atul got an important point.

  7. I totally agree with you! Hiding the protocol in some cases is confusing, inconsistent and doesn’t help anybody at all. You know, I suspect that in the near future, the address bar will disappear entirely, and be simply replaced by a non-editable field with the favicon + domain name only, arguing that for most users, it doesn’t matter on which page you are exactly within this domain.

  8. I agree that I don’t like this approach. About:config is my friend. Your post title is a tad sensationalistic though.

  9. I can see the point about how many people would benefit from having more confidence and knowledge about looking closely at a URL, understanding its construction and how to edit it. A great example is how you can often skip to deeper content rather than having to scroll page by page by editing a parameter number in a URL.

    But … all the examples I can think of are to do with later parts of the URL, and not the protocol part. It seems to me tidying this away doesn’t hinder learning, it helps learning – as it helps concentrate on the parts such users may actually benefit from paying more attention to.

  10. As a long time fan of URLs, I have noticed several things in recent Firefox (and not so recent all sorts of things that make me sad). For one, I’m typing something into the awesome bar. Historically, I’ve narrowed it to complete from only my bookmarks for a level of control, but due to random circumstances I now complete from frecency like everyone else.

    However….I browse a lot of web. It is hard to tell whether the hits I see are e.g. or . The titles look the same and the URLs aren’t displayed.

    So this is one problem. If we don’t have a reliable way of seeing *what we’re actually going to* not only is a browser hard to use, its also insecure.

    So are URLs user resources? Does it matter if I paste the link or ? If you’re of the position that URLs are just as meaningless as a (non-NULL) pointer value in C, then it doesn’t matter. They both go to the same place. If you are of the position that URLs can convey information, as I am, then the two are very different. The first you might be able to guess you are getting some microfiction and the second is just a bunch of garbage.

    I worry getting into a modal thought pattern about some of these things. Text search has proved a really successful way of doing a lot of things on the web. However, it is not the only way and it is often treated as if it is. Ten years from now, we’ll look at that and laugh. Will we look at the relegation of URLs as hash tags the same way?

    I have a dream of programming from my awesomebar, of using it to convey information. As such, URLs are vital for me. I think it is sad that we haven’t made smart bookmarks a user feature. They’re hidden, and annoying to use, etc. Why would you google for e.g. Marsupial when you just want the wikipedia article?

    wiki marsupial

    would take you right there, given a smart bookmark. This isn’t rocket science. An interface can be built which will lead to more innovation and a new capability for Firefox that other browsers lack. You can bet google is leveraging its hold over the web through their browser. Why not use our chief advantage — the openness of the web?

    Note that I’m criticizing Firefox a bit here, but of course Mozilla isn’t the worst offender by a long shot. It used to be I could send links directly from a Google search page. Well, now for reasons unknown to me, the links you actually get are internal links to google that redirect. URL shorteners have made the problem worse by encouraging the URLs are just hash tags mentality. People don’t read URLs anymore. Just click click click. Phishing season, anyone?

    Mostly, I want us, as Mozilla and as denizens of the Open Web to make a concious decision about this instead of making a decision by not making a decision. If we want to hide URLs, fine, lets do it, but lets do it across the board and make sure we understand the implications. You don’t get to see where links go. You don’t get to see what page you’re on. You don’t get to see what cookies are for sites. You just get to click and text search. That’s it. One of the things that I found compelling about Firefox versus other browsers for the times is the Firefox *showed me* the URL and *showed me* where links went.

    I realize that I’m a web programmer and care about these things more than, say, my mother. What I would actually advise is a way to have a URL-free mode and a URL-is-user-content mode. You ship the URL-free mode but enable “experts” to see URLs *everywhere* they want to see them. And speaking of things that are ugly that aren’t fixed….you’d need a preference to do this, right? Should I care about archaic preference names any more than I care about URLs? certainly ‘enableFoo’ is as meaningful as ‘tag.781278’, right?

  11. obscuring the URL’s seems like a dangerous move. It presents a vulnerability to phishing, at least as I understand the concept.

    If you have something that is designed to hide the URL unless x condition is met, it seems to be fairly trivial to trick the browser. I could be wrong though.

    I think the emphasis should be on educating people, not following Chrome, or listening only to the designers. Not trying to be offensive to desingers, but security seems to be an afterthought at best, and considered a trivial waste of time at worst. Which is the opposite of how it should be.

    I guess I’m odd in that I think that the whole tablet-look trend in desktops is simply horrible, its a desktop/laptop, it has a physical keyboard and mouse, make something that is designed for those features. At least make the interface that looks designed for a touch screen optional. I want to see the URL, because I want to know where I’m at, I don’t walk around the street blindfolded, why would I walk around the internet “streets” blindfolded? The probability of being hit by a car is still increased, even if its only a virtual one..

  12. My preference would be to just replace the protocol with a small logo inline with the text. People could recognize little icons better than random letters. Clicking it could show more info and connection information. (What’s the IP address of the site I’m on? I can’t tell.) As far as I know, HTTP does not have such a logo.

    And obviously fix the copy-paste bug.

  13. You are simply praising for a developer mode in browsers. The browser that developers need and the browser that everyone else need is not the same. The type of crowd the browser has reached is a lot wider than initially, and that we like it or not, people don’t want to know what hypertext is, they just want to communicate.

  14. It’s sad that such drastic (however small they may seem) changes get in without proper discussion, and with much contempt from the few Firefox developers that are behind it. It’s sad to be following Chrome in this, and it’s depressing that the Mozilla manifesto is again disattended… so much for the better web! 🙁

  15. I really love these thoughts. Previously, I hadn’t really considered the consequences of removing the ‘http’ from the address bar; from Chrome’s standpoint, it makes sense to highlight websites that are secure. But Firefox at least already does that, arguably in a more discoverable way with a big colorful block.

    Your post brought me back to 1998, and viewing page sources to figure out what the hell I should be writing. I guess I still do that sometimes. 🙂 So I agree: Leave the http!

Comments are closed.