The www. doesn't really make any sense. Showing different contents on xyz.com and www.xyz.com is IMO disastrous.

But it will be pretty hard to get rid of www. since a lot of us grew up with www.something.com.

If anything, we started getting rid of the 'www.' before the 'http: //'. :) This post is on 'blogoscoped.com', for example.

I agree!

www.example.com IS NOT THE SAME AS example.com

It doesn't make any sense to get rid of the www. http is the protocol, it's always there (even if you don't type it in). 'www' is just another subdomain (which happens to be widely used for web content).
http and www are fundamentally different things, there are no grounds for treating them the same.

Generally sites don't show different content on www vs. non-www; a good site has one simply 301-redirect to the other.

That said, the 'www' (or for that matter, any subdomain) has real usefulness due to an obscure part of how DNS works. For instance, if you do a "dig www.google.com. a" on some Unix-type box, you'll get a response including:

www.google.com. 132631 IN CNAME www.l.google.com.

The special bit here is CNAME, which is a domain alias. There's many reasons to want to be able to use a CNAME: for Google, it allows for location-sensitive DNS clusters to provide the actual address of "www.l.google.com."; for me, it allows me to point all the domains I host in my personal webspace to a single "master" domain that controls the actual "A" address record. In both cases (google and my sites), there is a www-less version of the domain that points to just a redirector--its sole purpose being to point your browser to the www.-prefixed version. It's also quite handy for people just starting out with running a webserver and using one of those dynamic IP services; a simple CNAME can point the much nicer looking "www.example.com" to "example.dynamicdnsprovider.com".

Whether it's useful in a given situation or not, CNAME has one flaw that prevents its use at the top level of a domain: it isn't allowed there per RFC, and will confuse many DNS clients if it is put there anyway. CNAME is only legal by itself, alongside no other DNS records (except for DNSSEC, but that's not important for this point). However, a top level of a domain such as "blogoscoped.com." always contains other records: SOA and NS at a bare minimum. So at the top level, no aliases are allowed; only actual addresses can be used.

Yes, you're probably more confused than clear on the reasoning by this point, but the "www." has more usefulness to many domains than just an acronym of "World Wide Web". :)

http;// and www are similar in that they are both things that get typed millions of times per day with no significant result. IT's not necessary.

We also need a new DNS system that goes gmail/google.com – that will also cut down a lot of RSI because once the browser sees 'gmail' it knows exactly what to automcomplete.

> http and www are fundamentally different things, there
> are no grounds for treating them the same.

They are completely different, we know that – but they are both often redundant to read or write for a casual user (which 99% of us are, I guess). While there is no reason we should just treat them the same, I think it is useful to think what speaks for or against making the URL bar more user friendly in any way possible – and getting rid of http:// may only be one (but not the only one) way to make that happen.

> Yes, you're probably more confused than clear on the
> reasoning by this point, but the "www." has more
> usefulness to many domains than just an acronym
> of "World Wide Web". :)

Maybe I didn't make my point very clear – I'm not saying the browser backend should get rid of the www. I've asked for a discussion about getting rid of this in the frontend, i.e. in the URL displayed to the user (just like Google still uses the http, of course, it just won't show it to the end user). So behind the scenes, if for instance a user enters "foobar.com" and your server redirects to "www.foobar.com" then that would still all happen, only the user wouldn't see it (the user would still know once they copy the URL, as then it would be "http: //www.foobar.com" which they're holding in their clipboard).

So the question is not about technical needs for this – the question is about whether or not users should be trained into looking at the www. and typing the www. (because even if you are usually not required to type it, casual users who are used to seeing the www. may believe they will also need to type it... and they may not know of any ctrl+return shortcuts either).

But just to emphasize, I'm merely opening up a discussion for this. I'll make up my mind if I hear more up or downsides :)

I could see it being a part of the URL that gets the "grey" treatment (there's a browser, whose name I won't mention to avoid nausea, that does this already). But in practice, you'll see a lot of people in the browser development side of the world get quite irate over the prospect of a displayed URL not representing the actual domain name being used for requests.

Chrome does strip the www. as part of the "short version" of the link-hover information shown in the lower left. It reappears, along with the middle abbreviated part of the URL, after hovering for a couple seconds.

> you'll see a lot of people in the browser
> development side of the world get quite irate
> over the prospect of a displayed URL not
> representing the actual domain name being
> used for requests

Ok, but what about the end user? "foobar.com" is less to read than "www.foobar.com" and still recognizable as web address (through the .com) and may thus speed up recognition by split seconds, nevermind if the site in question decided to redirect me from non-www to their www version (I do such a redirect to www too for some of my websites [though sometimes also from www to non-www]).

The next step, I suppose, is getting rid of the ".com" :)
It being so much of a default TLD, you could enter "google" and it would take you to http: //www.google.com but it would only display google. Go to google.de, and the ".de" – being an "exception" – would still display. Well, the problem here I guess is that once the .com is removed the URL becomes less recognizable as actually being the URL due to the missing of the .com "brand".

NOOOOOOOOOOOOOOOOOOOOOOOOOOOO!!!!
the www part is NOT always here and I already have to "help" people who can't reach the right website as they add it to places where they shouldn't, please don't make it worse!

Having a smart browser who can add common prefix and suffix to avoid typing them is fine. Hiding them for real would just increase common confusion...

While we're at it, why not get rid of %nonsense, and show actual characters in URLs (while still copying %... to clipboard, of course)? I think just replacing %20 with actual spaces would benefit casual users much more than getting rid of www, com and friends.

(And one more: once we realize that copying URL from omnibar can get you something you don't actually see, it is about time we started copying "&"s to clipboard as "&"s.:)

I know several websites which don't return the same content with and without the "www"...

BTW, this idea is stupid, because many people just type "example" in the address bar to go on "www.example.com", thanks to "I'm feeling lucky" service of their browser. So they never type in "www".

Both getting rid of http:// and www are stupid.
- you don't really need to type them anyway
- space on the address bar? with the current screens there's a lot of space available (screens are increasingly bigger and cheaper, and they're panoramic...)
- you still have to show if it's http or https (Google Chrome shows https:// but not http:// ... that looks bad and can even be confusing)
- there are other protocols like ftp:// , which is the reason they put the protocol in the url...
- most of the times you just click on links, or use a search engine, or some kind of autocompletion; youdon't even look at the address bar
To sum it up, the only advantage of this decision is aesthetic, and it's not so aesthetic anyway... it's just they have to keep offering new things, useful or not, but they have to offer changes and "improvements" if the don't want the user to switch to another browser.

TOMHTML You are right.. I too know several websites which don't return the same content with and without the "www"...

So I still want my sites to have www.

Getting rid of http in the browser makes sense, because it's the default in the browser, and its the browser's choosing to use that protocol when you don't explicitly provide one. I agree that removing the need to type "www." is a good one, but not one that should be done in the browser – site owners should be encouraged to do this through DNS.

The difference is that for HTTP the browser is hiding its own default decision, but with www. it'd be hiding someone else's decision as to what to call their website.

First, you will have to fix AppEngine:

http://code.google.com/intl/en-EN/appengine/kb/general.html#naked_domain

"
Q: I'd like to map my app to http://myurl.com (also known as a naked domain).
A: Due to recent changes, App Engine no longer supports mapping your app to a naked domain.
"

The same problem applies to Google Sites and Google Apps.

...

Avoid typing "http: //www." is good for usability. Hiding it, is BAD.
Try to explain future wannabe web developers: why they have to enter whole "http: //" thing in "a" tag.

Oops... Yours blog engine eat all "h t t p" strings too. Funny.
Now try guess my point of view, when you can't see everything.

Here is another argument against hiding "h t t p : / / w w w".

[edit: I added a space to prevent auto-linking. -Philipp]

Red's posts above illustrate why the real URL should be exposed. Otherwise, everything starts to feel non-deterministic and it dis-empowers the users.

But the web designer should always make things as simple as possible for the end user. All of my websites are on the bare domain (without www.), and have a redirect from the www version, so that it doesn't matter which one the user types.

Technically, Todd Vierling is right when he says that the Domain Name System offers better support for websites on a subdomain than for websites on the bare domain. In particular, the CNAME record lets you configure large sites for load balancing.

However, my sites aren't ever going to be big enough for it to make a difference, and the bare domain name works fine.

[edit: fixed reference to Red's posts]

Maybe that would be fine for the average internet user, as they wouldn't really know the difference and cannot care less about what protocol sub-domains are. And in the end, it's all about users.

Still, that sounds like an heresy to me.
I want to see what protocol I'm using, and on precisely what domain I am.

The idea is old: http://no-www.org/

Argos (a large UK retailer) used to force you to add www. to get to their site. I don't know why, I assume it was badly set up, but without it you got taken to a 404 page. This has been fixed now, but if the user hadn't been used to typing www. they may never have found the correct site.

As mentioned above there are still sites that do this (but for legitimate reasons).

> people who can't reach the right website as
> they add it to places where they shouldn't

Which actually sounds like a perfect example of why we shouldn't necessarily even require it – because it can confuse some people, as you say.

> So they never type in "www".

Tom, but they still *see* the "www". (Some may also type it.) In fact, many people also didn't type the "http: //" and yet Google decided to hide it, to maximize address bar readability/ make it more casual.

> there are other protocols like ftp:// , which
> is the reason they put the protocol in the url...

What's bad about just showing the protocol for rarer use cases like say ftp: // (which I'm happy to use in a browser every now and then, but which likely most people won't on a daily or even monthly basis)?

> I want to see what protocol I'm using

Rafi, so you won't be usiung the latest Google Chrome for that reason I guess?

> Red's posts above illustrate why the real URL should
> be exposed. Otherwise, everything starts to feel
> non-deterministic and it dis-empowers the users.

Actually, it just shows that this forum engine has a bug with auto-linking – it simply shouldn't try to auto-link a "naked" "http ://" string. When auto-linking works right, it always omits the http :// here in the forum and it doesn't hurt anyone (I find it increases readability).

> This has been fixed now, but if the user hadn't been
> used to typing www. they may never have found the
> correct site.

IMO if you try to get visitors to your website (which site doesn't?) it's a very bad idea to force someone to type the www.

If people starts to write only domain.com like google.com, there will be a lot of misunderstandings. People will write google.com even when they mean docs.google.com or calendar.google.com.

KMB, if someone today memorizes "docs.google.com" why do you think they would be less likely to memorize that once "google.com" were to become the main display? Could you explain a bit?

By the way, I want to emphasize once more: I'm not pro or con "www" hiding. I was just hoping to find some pros and cons by you guys, to get this thought through a bit by us...

If you ask a forum full of tech-minded people about whether to simplify things, they don't like the idea. But of course for most users the "www" is just an extra bit of meaningless gibberish, and it could be removed.

I never type "www" these days, but it's annoying when I want to access one of the few remaining websites which display a 404 error without it (e.g. my old university: http: //ed.ac.uk does not redirect to http: //www.ed.ac.uk, but it should do).

I think this discussion is missing point a little.

It's not about using "www" domains. They are unnecessary today..
It's not about typing "www", every 5min.
It's about hiding "www" and "http" prefixes.

In my opinion browser should show full domain name.
(browsers are for reading information, not "censoring").

btw.
thanks for link to no-www.org

How about a slight change?

If the URL starts 'http:', remove that, leaving the '//' intact.
If the shorter URL now starts '//www.', remove that.

For most sites this will have the same effect that you described.

However, occasionally you will see '//domain.com', which is an indication that the full URL does not contain 'www.'.

For example, my domain might contain two sites 'site1.domain.com' and 'site2.domain.com' which would appear unambiguously as '//site1.domain.com' and '//site2.domain.com'.

> browsers are for reading information, not "censoring"

I guess then the question is: which useful information on average does the "www" bit hold (to the average user)?

lets not completely confuse users and businesses by having some users put in the full www. and others only putting in the domain name. some sites do use the www for the public side of their business and sometimes the domain only for a private, more secure section. www is a subdomain. if an entity chooses to use it then it should be visible. we are trying to teach users to make sure they know the domain they are accessing and to them if they type www.whatever they want it to display so they know it is where they want to go. don't mess with it.

why not just create an interface where you just go to a page and then type in a keyword to go to the site.. oh wait.. AOL did that and see what that got them. leave the Internet alone.. we can make it pretty, but let's not change out the conventions.

the www is utterly useless and should be ditched. Sites which still require it really suck because its just a subdomain of something (and the world wide web is a strange term for a subdomain.
Rule should be.
x.com = Public site
y.x.com = special site, different format or for internal use
...and the www should be fased out.
(...unfortunately their are some sites which are only accessible with the www because the main domain is not used in dns entries).

Thumbs up! Should this even be a question ? What´s the real point in using it, other that serving as a efficient time-waster, for some noobs of course ? Just the name of the website and the TLD ought to be enough, but this does raise a question in my head. What if this could spell doom to WWW in the minds of people ? Of course that´s silly, but think...

Jim Perry
I do not agree. Most users are used to go to x.com and do not use www.x.com but admins have to configure x.com and www.x.com with the same dns entries which is a waste of time.
...and I know no company which uses www for their public site and the main domain for internal purposes only.

The real focus should be getting site creators to stop using 'www' as a subdomain – www is a joke that isn't funny anymore: a 9 syllable abbreviation for a 3 syllable phrase... Bleah.

I say lose it. Google should start a campaign to get sites to drop it.

How about a poll?

Im for dumping the www... yes, youve been using it since you were a kid...yes things change and it can be awkward but its the right thing to do.

getting rid of the http was a great idea.. but judging by the bug report entry on on code.google.com you would think someone murdered their unborn child. think of the man hours training people to stop useing the www would save..

The main public face of a domain should be the regular domain name. use a subdomain for your intranet. I dont think all subdomains should be not displayed.. just www..

Philipp Lenssen

> I guess then the question is: which useful information on average does the "www" bit hold (to the average user)?

None.

But it's not the point. It's not the browser job to hide "www" prefix.
Why would you want special treatment for this one prefix?

Websites should simply stop using it. Google have position to promote this behavior.

Philipp wrote:

> I guess then the question is: which useful information on average
> does the "www" bit hold (to the average user)?

The "www" is needed to uniquely identify the website we are visiting.

The situation with "www" is quite different from that with "http ://".

If we adopt the convention that we will drop the most common protocol (http://) then we still have a "one-to-one mapping" between the URL and its representation in the user interface of the browser.

But if we drop the most common subdomain (www.) we no longer have a "one-to-one mapping". If the browser shows "foo.com" we no longer know whether we are visiting "foo.com" or "www.foo.com".

[added space ~ hebbet]

To everyone suggesting that we should encourage sites to drop "www" as a subdomain: if browsers simply stopped displaying the "www", then probably companies would quickly abandon use of the subdomain. Anyone publicising their website with "www" in front would look old-fashioned.

I'd go so far as to the say the address bar should serve two purposes:

1) a place to identify where you want to go – either be entering the IP address directly, entering all or part of the URL, looking it up in your history, or entering a search string

2) the page title once you've arrived – i.e.: DO NOT display the URL but the TITLE!!!

I'm never going to type in the entire URL of a page I want to get back to. At most I'll type the site's URL, without the www and then navigate to the page I want. Most likely I'll find what I'm looking for in my history.

Well, I'm not reading *all* these comments, so I may repeat something already said:
1) www is a subdomain and where it points depends on the configuration of each domain.
2) Google uses www.

Rob Fuller:
3) Is not "old fashioned", buddy. Is a subdomain, and it points to a particular computer, as it would www2.example.com or bob.example.com.
We all want beautiful URLs, but there's a reason for them to have this syntax.

4) Is sad, but most of the people relate www with "internet", and if you write just "example.net" they won't get it is a URL (because .net is not as known as .com, and because there's no www at the beginning)

schultzter: I care more about the URL, rather than the title; think about all the phishing possibilities you could have if the URL changed to the title... u_U

Rob Filler
quite the contrary. When you hide the "www", most will lose the motivation to bother with redirects.

The next time someone asks why he must enter "www", he will hear the reply "install a modern browser" ..

In this way, "www", will remain forever.

Red

Maybe you're right, but in any case it wouldn't matter whether www were technically part of the URL or not. Normal users (like myself) would never see it.

Zim

But I'm talking about presentation, not the technical back-end. Users shouldn't need to know which computer the site is hosted on at your end.

I think that you are right that the main value of "www" has been to identify website addresses in publicity materials. But I'm seeing more and more companies omitting it now (and not just with .com addresses), so it won't be so well recognised in the future.

I actually agree with schultzer that in the future the browser won't normally show a URL at all, but instead something like a website's title or breadcrumbs. I see a move towards that in the way that Google search results already sometimes display breadcrumbs instead of a URL. The URL bar would be an extra toolbar, which advanced users could display if/when they needed it.

How many internet users even now would be able to recognise a potential phishing problem by looking at the URL? Very few. There are better ways to deal with security threats – such as the way that Chrome produces warnings.

They could offer both: with http: //(www) or without: You click in the address bar, the URL expands to the prefixed-version. When the address bar is not in focus, shrink it for readability.

The best of both worlds.
Would people still be confused? Definitely.
But they're the same bunch who'd be confused, no matter what.

http://usps.gov/ (United States Postal Service) is one of the Government sites that constantly tricks me up by not mapping the no www to the www site. They do map the usps.com site but its just idiotic to not map the www. Most often its US Government sites that have this problem, and once and a while some large old corporation.

I personally never type the www unless the site bombs out and will not load.

If they are going to start hiding it I might prefer that they try the www. domain when the main domain doesn't load. Some browsers already do something like this where they will try coke.com if you just type coke in the url field.

no-www.org provides a server to test the status of www subdomains.
http://no-www.org/verify.php

Yes, I agree entirely. I often miss out the www part anyway. Sometimes that gets me a blank page, but I get to the right destination 95% of the time.

The thing is, not all sites work without www.

A local bank site has this address: "www.bancomercantil.com". If you try going to the site without using www you won't get anywhere. I'm not sure why they're doing that, they may not even know, but still happens to a lot of sites.

So that's my opinion: all sites should get rid of the www :D

I think them getting rid of the http:// was already stupid. They should keep the URL the same as what was entered.

Hi Philipp,

I'm the guy who designed the omnibox and one of the folks who implemented scheme hiding.

We won't (and can't) ever remove "www." because that changes the meaning of the entered text. Similarly, we won't remove ".com" as you suggest in a followup comment. "google" is a search term or valid intranet computer name, and isn't the same thing as "google.com".

We don't play tricks like remember the "real" meaning of the displayed text and use it under-the-hood, because it becomes impossible in such cases to discern the user's intent when the displayed text is edited slightly (without being completely replaced).

> We won't (and can't) ever remove "www." because that changes
> the meaning of the entered text.

Actually, if the user entered "www", you could leave it as is – but if the user enters "foo.com" and foo.com redirects to www.foo.com you could (and this is not a suggestion, this is the hypothetical basis for a discussion I'm inviting others to join here) omit the www in the display. In other words, people who don't care about entering "www" wouldn't be pushed to thinking about it. I agree it would not be very nice to *remove* a "www" once actively entered by the user, as that would perhaps be too much overriding of user intent, I don't know*. You also don't hide the "https ://" in Chrome when I enter it, yet you hide the "http: //" when I enter it (because the latter is considered "default")... so people who care to enter https are fully respected in doing so, nothing changed with that even with your http ommission.

(*admittedly, you even hide the http :// when I actively/ manually entered it)

> "google" is a search term or valid intranet computer
> name, and isn't the same thing as "google.com".

("google" being both intrant name and search term would actually mean you would absolutely need the "http ://" in the display to differentiate between the two during typing-and-auto-completion-selection... yet you managed to differentiate in another way, namely by using two different icons [magnifying glass vs earth icon]. But that's a whole different discussion, that .com thing...)

> We don't play tricks like remember the "real" meaning
> of the displayed text

(You do remember something though: When I enter a local host domain name [i.e. coverbrowser, no .com, which I set up on my local machine] then I'm getting a Google search for the term, but there will be a Chrome question asking me "Did you mean to go to http: //coverbrowser/?" If I say yes, Chrome remembers it, and next time I enter the exact same thing into the bar – "coverbrowser" – it will understand to go to http: //coverbrowser/ on my local machine.)

All in all, worth noting that display and text entered are two different things. Just think of the Google Webmaster Console, which lets admins define a setting which says "Display URLs as example.com" (vs "Display URLs as www.example.com") even when that domain actually "forces" the www through a htaccess redirect from non-www to www. I don't know how Google actually treats this setting once enabled and I'm NOT saying that a browser URL address bar is comparable or even mildly the same thing, by the way (i.e. I'm sure there's plenty of differences between address bar display and webmaster-configured result display, no arguing about that) – just pointing out a case of display vs real URL.

But again: I'm not suggesting anything like leaving away the "www" or the "com" – rather, I am curious to hear good reasons for or against such omissions, so I posed this as an open "maybe" question in the post. I'm completely open on this and rather wanted to hear arguments pro/ con.

[While I am talking to a person responsible for Chrome's cool address box: I found that the earth icon that is now used instead of "http: //" draws too much attention to it, perhaps it has a too-high "complexity" (or is not transparent enough?) in comparison to the simpler other icons? Perhaps displaying an earth is also just to content/ expression heavy in a browser that tries to have its interface "disappear" to make place for the website. Dunno (and dunno if there's a better alternative). I find Chrome generally does the most excellent job of all browsers to "disappear". Perhaps there just isn't a better icon or gray/transparency-level or level of abstraction than the current earth you're using.]

Now, if we have to boil the www issue down into a single question, perhaps it's this: What practically speaking would be bad for users if a browser bar would, when the user enters "foo.com" and foo.com redirects to www.foo.com, to not show the "www." in the address bar but just show "[earth icon] foo.com"? (One would still be able to see the full address one ended up on if needed – i.e. "http: //www.foo.com" – say by clicking on the earth icon next to it, and if one copies the address bar to the clipboard it would still be the full URL in the clipboard, etc.) I'm really interested to hear all the con arguments.

> and if one copies the address bar to the clipboard it would
> still be the full URL in the clipboard

I don't like the idea that if I highlight some text and copy it, the clipboard gets something different from what I highlighted.

> I don't like the idea that if I highlight some text and
> copy it, the clipboard gets something different from
> what I highlighted.

Fair enough, but that's exactly what Google Chrome already does (perhaps you're not a big fan of their bar or their http omission decision though... personally I really like it). If the Chrome bar displays

www.google.com

and you copy it, your clipboard will hold

http: //www.google.com/

They omitted both the http and the ending slash from showing (to make it more readable and less tech gibberish for casual users I guess), but then put it both in the clipboard.

To my mind, putting extra text in the clipboard is as bad as those websites that use JavaScript to append a copyright warning to any text that you highlight and copy.

So if I was writing an article about domain names, I might say "Google search uses the domain name www.google.com", and instead of typing that domain name I would copy-and-paste.

Ugh, my article then ends up with a URL rather than a domain name.

It's as bad as when Windows dropped filename extensions from the Explorer by default. Sure, it gives an illusion of simplicity, but it creates problems in a number of edge cases.

Hiding parts of the URL (protocol or other) is creating stupid new users that no longer will understand the meaning of an URL as identifying how and what will be retrieved by the browser.

These users will become even more prone to scams as the result of malicious redirections, as they lose the ability to even get a clue from the URL that could indicate something is going wrong.

IMHO Chrome (and other browsers) should not mess with the URL – period. They could colour code the protocol like Chrome already does for https (green) and https URLs with http parts on the page (red) but they should never remove the protocol part.

As for the www part, the same applies, never hide what is actually part of the URL, colour it if you must but don't remove it.

And if the browser maker really think they know better and that such should be for the average user, they should always provide an option to undo their decision for those users that don't agree with the simplification.

Ahab, hiding the www could actually hide one source of scamming potential – think of domains like wwwpaypal.com (note missing dot). Even a subdomain like vvvvv.google.com could create scamming confusing (in case you didn't notice, I typed "v"s instead of "w" for the www – though admittedly I don't know how one could create a scam with such a trick in the subdomain). Or could you explain how hiding the www could be used in scams? But in general perhaps everything that simplifies the URL could make it easier to read, to the point where even a novice user might understand that something's wrong when something's wrong... because you'd have to change the main brand (say "google.com") in the URL to get a scam working. But of course, it's all very hard to tell for sure...

a lot of websites simply don't work when you leave www. out of the domain.

Philip,
I simply think that hiding parts of an URL is not adding anything for the user. It only makes stuff confusing. Take a simple example. Suppose you are a relative new user of internet services (because you a young, were never interested before, or just a late starter. And you get from your internet provider the instruction to type in http:// www.yourISP.net to find the login for your account. The instruction – which is aimed at inexperienced users – tells you to type in exactly that URL and that you to be sure you are at the right address that you should see http:// www.yourISP.net in your browser bar.
Now what good would it do to this new internet user to hide parts of the URL to him/her? IMHO it will only confuse that user as (s)he tries to follow instructions to the letter when the browser does not show what this user expects.

Should that ISP make separate instructions for each make of browser? Or should the browser makers realize that it's better not to hide stuff that other browsers show, or rather that are just part of the URL?

It would be like not repeating a street's name at each street crossing, just because the street name was visible at the start (and the end) of the road. Repeating the street name is not necessary if you exactly know in which street you are, but it surely helps when you are in doubt.

Likewise the URL should not be changed by the browser. Not even by hiding the thought of as familiar (but not redundant) parts.

It w***d be as u****l as l*****g o*t p****s of w***s b*****e t*e w***s a*e f******r to t*e r****r. It would make sentences more terse to read and make books a lot smaller but it would leave the same reader confused about how to interpret some of what was written no matter how familiar these words would be.

Note: Where you see yourISP.net please read h t t p : / / w w w . y o u r I S P . n e t (without the spaces).

Get rid of it. Plain and simple.

> And you get from your internet provider the instruction to
> type in http:// www.yourISP.net to find the login for your
> account. The instruction – which is aimed at inexperienced
> users – tells you to type in exactly that URL and that you to
> be sure you are at the right address that you should see
> http:// www.yourISP.net in your browser bar.

What's interesting here... I think your example is a perfect example of why we should not concern most end users with stuff like "http: //". What good does it do over the phone to give an instruction like "please enter age tea tea pee colon slash slash" – "What's that?" – "... colon slash slash ..." – "That's two slash right?" – "yeah right, slash slash then double you double you double you dot..."

It's so much easier to tell the user: "Please enter google dot com, and tell me what you see". If at this point you wanted to confirm what's in the address bar you still could (it would read google dot com, and you could also ask "is there a globe icon next to it" if you wanted to make extra sure, tho probably there's no need for that).

> Should that ISP make separate instructions for each
> make of browser?

I think already they will absolutel have to if their intent is to get the instructions right word- and pixel-perfect. How else are you going to explain to the user a simple thing like "Click File -> Options" in the menu, when the options may be accessed in completely different ways already in different browsers, and may have completely different naming and hierarchy?

> Repeating the street name is not necessary if you
> exactly know in which street you are, but it surely helps
> when you are in doubt.

To follow the street analogy, I think it's a good one, because we can think of
"Something Rd."
as
"Something.com"
which may be an indicator to not get rid of the ".com" ending in the display as it can serve as part of a "URL recognition brand" thing.

> It would make sentences more terse to read and make
> books a lot smaller but it would leave the same reader
> confused about how to interpret some of what was written
> no matter how familiar these words would be.

Tutorials and technical guidelines merely aimed to get a point across should be terse. Other books may be longer, but we read them for fun – they are content. The browser itself however is not there to be fun to be read, it is there to get out of the way IMO and let users see and do what they actually wanted to see and do... i.e. read websites.

OK let's assume the http:// – protocol – part in a URL is to be considered redundant.
Then in order not to confuse the user any more let's get rid of the browser showing any protocol.
No more https:// no more ftp:// no more mailto: (although technically that may be not a browser supported protocol, at least not in some browsers).
   Let's get rid of all protocols or URI schemes *) because they offer no useful information to the user that is already overwhelmed with so much information.

While we're on it, let's do away with the entire URL and replace it by a user friendly description; no longer h t t p : / / blogoscoped.com/forum/174339.html but instead the much more meaningful 'Blogoscoped: Google Chrome already got rid of the h t t p : / / , maybe time to get rid of the www. too?'. Because we don't want to be bothered with URLs. That description should be enough to bookmark and if we would like to find that exact web page back search engines are smart enough to find it.

Or do we? Do we want to abandon any and all control of what we want to bookmark and store and refer back to, or which we want to forward to others? Do we want to have it all dumbed down and non-technical and easy and so utterly out of our control because there is always a browser or operating system or search engine that will cocoon us in and make sure that we get what we want. Do we really want that? Do we really want everything that even only marginally may appear as technical or complex hidden away?

Please don't get me wrong, I'm all for progress and innovation. But I'm also convinced that before you do away with the old you should be 100% sure that what is new also is better, really better. If it ain't broke, don't fix it.

(And to go on with this rant, but then about Google, they – the Google mangers and engineers – are continuously tinkering with their products. But not every change is a change for the better. Some changes are just changes for the sake of change, only because change would indicate they are being actively involved with a product. And at the same time this leaves the bug and problems that already exist for ages in the applications – because visible change is deemed more important than making an application actually better, and stable, on the inside.)

*) http://www.iana.org/assignments/uri-schemes.html

> No more https:// no more ftp:// no more
> mailto: (although technically that may be
> not a browser supported protocol, at least not in
> some browsers).

I think it's important to show the https, either by icon or perhaps better still by using the actual letters, because only http could be considered "default" and thus be omitted.

> Do we really want everything that even only marginally
> may appear as technical or complex hidden away?

As easy as possible, yes, but not easier. Your examples of getting rid of the complete URL add problems in communicating the URL or knowing just where you are – something which doesn't happen when you merely omit the http part (or, some may argue, omit the www part).

> If it ain't broke, don't fix it.

Interestingly enough even the web's inventor, Tim Berners-Lee, admitted that the whole colon double slash thing wasn't the best naming convention, and one of his regrets (I'm paraphrasing, but Tim said something to this regard... and perhaps he was half-joking, but that still leaves half-honesty). Perhaps it was indeed usability-broken all along for casual users.

IMO it's not black or white – it's how to make the experience more user-friendly on a gray scale. It's not about "They removed one button, what, do they want to get rid of ALL buttons?" No, but maybe using 5 browser buttons is more friendly than using 25. I hope that more people will consider casual users and everyday usability because often, we can combine what's technically right with what's easily usable. I recently saw someone say "Of course using technology is complicated and getting more complex, after all, technology also offers more and more and is getting more powerful!" But I think no – I look at something like the iPhone, which is much easier to use than most of my phones from years ago, and yet it has many more features than my old phones. With the right designer and developer philosophy you can achieve greater features while still simplifying the interface.

The iPhone, btw, also omits the http: // in its display (but when you tap the address bar to enter something in Safari mobile, it reappears). The iPhone's keyboard even offers a ".com" key, which technically speaking is also "wrong" because .com is "just another top-level domain" so they'd be required to offer .de, .net, .whatnot at the same time. They even ignore your dot if you type "google." then hit ".com" (which technically speaking surely must result in "google..com", or else, user input would have been ignored). And yet, it becomes a great time saver in everyday usage, simply because .com is the average, the most popular, the default for many users (just like http, and www, is).

Zim You do have a valid point with regards to phishing – but the reason phishing works today is because people don't check the URL in the first place. And in the end, all modern browsers, anti-virus software, and even DNS services (like OpenDNS and Google's DNS) check and warn of phishing sites. I think I'd feel safe seeing the title instead of the URL once I got to the site.