Ever since the first traditional media ads began to appear with Web site addresses in them nearly ten years ago, people have been screwing them up. And unless you just got on the Internet yesterday, you’ve almost certainly done the same, and know many people who have as well.
Almost no one knew what http://www.pepsi.com/ meant at the bottom of that TV ad. Those who did almost certainly had mixed feelings about it. Until 1994 commercial activity on the Internet had been prohibited; when it was lifted many people feared the inevitable commercialization of the net. In hindsight, they were right. Now commerce runs rampant. Overall, though, this isn’t a bad thing. Sure, there are banner ads and spam, but the Internet also brought us Linux and eBay.
The thing is, in 1996 almost everyone on the Internet knew how to type in a URL. Now, in 2005, almost no one knows. They guess, and the browser makes its best attempt to find what the user had intended. Back then, virtually everyone dutifully typed http://www.pepsi.com/ into Mosaic or Netscape to see just what it was that Pepsi had put online. These days, many people will just type in www.pepsi.com or even pepsi.com or, believe it or not, just pepsi.
More people than you realize will take a URL, go to their favorite search engine, and type the URL into the search engine’s search field, never realizing they can actually edit the contents of the address bar above, or perhaps not even noticing it. I’ve even seen a few pathological cases where, given a URL, they will type www.google.com or another search engine into the address bar, and then type the URL they actually want to go to into Google’s search field!
I paint this bleak picture primarily for the benefit of Internet veterans who have been around the block a few times and actually understand a fairly good deal about how the technology underlying the Internet works, and who don’t deal with “normal” users on a regular basis. If my description of “normal” users above surprised, shocked or disappointed you, you’re the target audience.
What does all this have to do with “www.” in URLs?
Remember that some people always use it, some don’t, and some only use it if it was in the URL they were given. The source of this confusion is the simple fact that users today don’t understand two things. First, they don’t understand why the www. is (or isn’t) there, and second, they don’t understand that the Web is not the whole of the Internet.
The main reason I argue for leaving the www. in URLs is that it serves as a gentle reminder that there are other services than the Web on the Internet. Some of these, such as FTP and DNS, users typically use transparently without even realizing it. Others, such as e-mail, users access through separate applications. Even so, I know of many users who will claim with a straight face that e-mail is not part of the Internet.
The ultimate goal here is to reach at least a few of these people and turn some of the lights on in their heads.
Please note that I am in complete agreement with the no-www people that a domain’s main Web site should be accessible through both domain.com and www.domain.com. I argue, however, that www.domain.com is the preferred URL and that users going to domain.com should be redirected to www.domain.com.
http://www.je-not-care.com
One of my professors in college used to make the same argument. However, forcing the www seems to sacrifice usability to try to teach the user something they probably won’t learn anyway.
Also, your argument that users will notice the different services (ftp, dns, etc.) is just as confusing. Are http://ftp.something.com and http://www.something.com the same server or different? Does a server have to start with ‘www.’ to work in a web broswer?
Because there is no standard answer to these questions, trying to enforce anything is futile. This is why we have ports, after all. Ports are hidden from most users simply because they are standardized. The standard lets us hide them, not show them.
Unlike FTP, WWW is not a protocol. HTTP is.
If you want a gentle reminder of the different protocols, argue for leaving in “http://”, and “ftp://”, instead of “www.” You can have a “www” prefix in a FTP URI just as good as in an HTTP URI–both URIs are part of the www (which is why I think the thing is redundant).
“The ultimate goal here is to reach at least a few of these people and turn some of the lights on in their heads.”
My ultimate goal is to make things easy for the user, and I think leaving out www. certainly makes things easier. How often have I tried to send someone to a url that includes a subdomain and got complaints for not reaching the site because they had typed www. in front of it?
I think that omitting the www. part saves time (money? as in print or redirects) and confusion (I personally hate it if I type an address without www. I get no response or no redirect). I see no disadvantage in teaching people to omit the www. part, therefore I redirect all www. requests to the non www. url for my site. Keeping www. for historical reasons makes no sense to me, sorry.
I would have to agree that the real focus should be on the http://
I think if more people paid attention to that, they might also learn that https:// signifies an SSL connection, and wouldn’t go around dumping usernames and passwords into as many phishing sites.
I’d have to agree with Daniel above. Looking at the web from a user-interface perspective, the users do not (and really, shouldn’t have to) know about www. or ftp. or http:// or the port numbers or anything. Ideally, the user would just type what he’s looking for into the address bar and he or she would get where they wanted to go. That’s what I love about the way firefox’s address bar works — it’s insanely smart. If you don’t specify http://, it’s assumed, because 90 percent of the time it’s http. If you don’t specify www, it’s assumed. If you don’t specify anything and just put “UC Berkeley”, then it uses Google I’m feeling lucky to take you there, if you’re lucky.
No, it doesn’t teach users how to conform to the tech standards of the internet, nor is it a “gentle reminder that there are other services than the Web on the Internet” . It’s just useful — it doesn’t require the users conform to it, it conforms to the user. For those of us who need to use other protocols, subdomains, ports, etc. etc, then we know how, and it’s still just as possible. If it’s a good piece of software, everyone else should be able to type what they logically think should go in the address box, and it should do what they expect it to. That’s good user interface design.
And to Peter — a small “s” in the URL is not a good indicator of security in a user interface. IE and Firefox now handle this fine by showing a lock and turning the address bar a goldenrod color to show you that you’re in a secure site. Those are user interface elements; an “s” is not, and means nothing (and always will) to the great majority of users.
I agree with Daniel on this one.
WWW isn’t a protocol in itself, it’s HTTP that’s the protocol that the WWW uses. Therefore I believe that the focus should be on the http://, not the http://www., since it’s the http:// that symbolizes an HTTP connection.
And, also, not every site has http://ftp.example.com, for example, my site, shorty114.net, can be accessed via http://shorty114.net/, which is the web server, or ftp://shorty114.net/, which is the FTP server. In none of these is there a http://www./ftp. subdomain, because it’s the http:// or ftp:// part that shows the type of connection.
Should we add an “email.” prefix to our email domains? Or maybe imap., pop3., smtp.,?
I don’t think anyone’s claiming that www’ is intended to describe the protocol, but rather to describe the service. Think of webmail; would you complain that using ‘mail.domain.tld’ or ‘webmail.domain.tld’ for that are incorrect since ‘mail’ and ‘webmail’ aren’t protocols?
We’re talking about giving things a name here. What is a proper name for a webserver?
The example above stating that ftp://example.com and http://example.com are merely enough to distinguish protocols is technically correct. But it forces to combine FTP and HTTP on the same machine (or at least to have one single machine to connect to, whatever happens behind it).
Although this doesn’t matter for most domains, which can afford shared hosting together with a thousand other domains concerning performance and bandwidth, whenever you have to take the step to install more than one server to serve your visitors, you certainly end up having to establish more than one name for your servers.
What does happen if the above example.com website has to split FTP and HTTP onto different servers? Links break! Because either FTP or HTTP have to change servernames, so you end up having ftp://server1.example.com and http://server2.example.com.
Of cource you can try to save the situation by leaving HTTP traffic at http://example.com, but your FTP server has to change it’s name.
Conclusion: It is better to plan ahead and provide different names for things which can technically be different things, and then end up pointing all the name to only one server, instead of establishing only one name, but afterwards realizing that this path leads to a dead end.
Redirecting (as a convenience for lazy users) http://example.com to http://www.example.com is fine with me. Doing it the other way round, either.
Michael, you’ve got to come up with a better argument than this:
HTTP is a protocol. http://ioerror.us and ftp://ioerror.us are equally descriptive and neither is ambiguous. A subdomain called “www” is superfluous and it only serves to confuse people.
I think you just made my argument for me, and betrayed your lack of knowledge of how the Internet works.
First off, www is not a “subdomain,” it’s a “host.” Second, how do you intend to disambiguate the HTTP server and the FTP server, which may well be located on different hosts?
To Andy Skelton: What’s confusing is getting redirected to after explicitly typing in .
To shorty114: What do you plan to do when you want to serve HTTP and FTP from separate machines? You can only give example.com *one* IP address, meaning you are then stuck using the same computer to serve HTTP and FTP. This is why you have hostnames under the domain name.
You out up a really good argument, but it is wrong to keep www. because it will confuse users into thinking that it is some sort of protocol (if they even realize that the Internet isn’t just Internet Explorer, sad but true…). I know you know what a subdomain is, well www. is a completely unneeded subdomain and therefore should not exist. WWW and www. are not the same, even if they have the same acronym meaning, one is just a meaningless subdomain and the other is a an application.
I don’t expect websites to start with www and am perfectly happy to see URLs such as http://phatblog.example.com/ but I don’t like the idea of http://example.com/. The main reason is that although the DNS system allows for example.com to be both a host and a domain, it is confusing and IMHO best avoided.
And what if I have two websites at example.com, http://www1.example.com/ and http://www2.example.com/? Even if I can overcome my reluctance and add an A record or a CNAME record for the bare domain to the example.com DNS server, which site should http://example.com/ refer to? Or to put it another way, why should one assume that there is a particular web server associated with example.com? There can be any number of hosts at example.com any of which could be running webservers on one or more ports.
Pingback: Isulong SEOPH / Isulong SEOPH URLs: No-WWW or Yes-WWW?
yesterday i was on the no-www side, but today after checking out the sites like google, cnn, w3c i decided, that there’s no space for some no-www experiments and sticked with the no-www to yes-www redirection… interesting that these big players are using 302 codes instead of 301 which i would expect.
www. is only a host (A record) on DNS, even a something.com can be a host, not a must to have http://www.something.com, and browser opening the host with http protocol.
Pingback: Next Generation Internet » no-WWW or yes-WWW
It is my view that the use of ‘www’ is no longer necessary, difficult to say and repetitive.
Pingback: You Must Decide! www.domain.com Or domain.com? at TenthOfMarch.com
One major point being missed here to do with SEO:
While saying “pick one and stick to it” is technically fine, we need to remember that users won’t necessarily. Why does this matter if we have a redirect? Well, from our experience of running a large site, we liked without the www as it was quicker to type and used less space in our logo, so we submitted without the www to all the search engines. The problem was that just about everyone who was giving a link to us ended up linking to http://www.ourdomainname.com!
Actually, I do not quite agree with the reasons behind yes-www but still I believe it is better than no-www. I am thinking of this without caring about protocols and other technical aspects. I am basing solely on the fact that on advertises it is better to see (more pleasing to the eye) “www.example.com” than “http://example.com”.
The first one is just more straight and simple. One can argue that on advertises, tv spots, posters, etc, the website should be written just “example.com”. But if you are using more obscure domains (ending with .it or whatever different from .com) there is no way of clearly indicating you are talking about a website. Thus, you need a prefix. And www. is easier to write and remember than http://.
What I actually believe is confusing (at least to the common user who just wants to visit a website) is the “http://” protocol. World Wide Web (www) is a better acronym for the regular user, and for any possible ad it just looks better.
I agree with the pro-www position; looks a lot better and makes sense. Why’d there be only one host on mydomain.com? There could be many. Just say mail.google.com, groups.google.com, images.google.com and so on.
You might expect the www to be the “front page”, or the www computer. But I have domains without a web page at all, the internet is not only the web. So I have no www in those cases, but might get one day.. You don’t know what machines will be connected in a few years when you get the domain.
Another argument is don’t fix it if it’s not broken..
Pingback: Yes WWW or No WWW | House Cat Productions
The pro-www arguments are dumb. The “www” aren’t needed at all.
If I type http://example.com, of course I’ll go to the home page. It doesn’t mean I can’t use http://mail.example.com or http://downloads.example.com.
“www” just make it hard for you to spell a site and require some more bytes. May be nothing for you but I guess Google would save some GB if sites haven’t “www”.
If sites uses both “www” and no-www, people will access them same way, although if you use some scripts like phpBB and SMF you can make an mess with cookies, as users logoff if acessing the other way.
There’s absolute no advantage to keep using “www”. Just drop it, for Pete sake!
If you look at mail.google.com, groups.google.com, images.google.com and so on, a plain “google.com” looks like a car without tires.
I can’t think why the hell it is easier to spell http:// than www. In any matter, drop http://
https, ftp and others could be added as an “advance” part.
Sorry. I didn’t meant http://www, just “w w w”. But it added the http thing.
To all you nay sayers:
Copy and paste the following into your email program, send it to yourself, and see which produces the clickable link. Anyone in the know knows how pattern matching works. Not only email programs are affected by pattern matching of this type.
*****
http://www.yes-www.com
no-www.org
*****
In the above i typed www+yes-www+com (using dots instead of pluses) and it got converted to http://www.yes-www.com by a pattern matching script in the comment handler. There it goes again!
Oh it’s an efficiency dictate.
An interesting example of the www/no-www phenomenon was for a time http://getfirefox.com – I tried to download Firefox and it gave me an http error page of some kind.
a www is easier to remember and it’s very easy to write a regular expression to highlight www-ified links. These in themselves could be considered good, strong arguments for keeping the www subdomain.
Yes, yes, it’s a host. But syntactically “google” is a second-level domain, so “www” becomes a third level domain – beneath the identifying domain hence “sub” – so it’s a subdomain like it or not, regardless or whatever else it’s called.
Anyway my personal story…… I had a couple of subdomains outside of my control at one point and for security reasons determined that my cookies should not work across subdomains. For this reason, I simply could not have both www and no-www friendliness.
I 301 now to the proper site, for what it’s worth, if someone “gets it wrong” I chose to go with no-www, because I find it easier to type. Either way, some peopel will get it “right” and some will get it “wrong”.
Personally I type in only “google.com” every time. I’ve looked at more web pages than hot dinners, if I had typed in the www every time I’d visited google or altavista I’d probably have stabbed someone by now.
But that’s just me. Definitely, it’s easier to communicate urls with the www – whether over the phone or in an email, it’s much easier to parse. We’re used to hearing it. You know as soon as you hear it that there’s a url coming, even if you don’t know what a url is called.
To those people who claim that having ftp and web on the same domain will hurt expansion, that’s just silly.
- load balancing comes into it at some point
- 301 redirects
- moving services as demand grows, and just mirroring them at the original site
To me, the protocol bit is the important part. http: … It’s insanely hard to remember as a computer newbie – I remember back in the day I had to go get a cereal box or a commercial softdrink giant’s product every time I wanted to type a url.
I was a Windows user – all the directory separators are backwards in Windows. Nutscape and Virus Explorer back then couldn’t handle websites like this, and neither could IIS of the day IIRC:
http://www.somewebsite.com\somedirectory\somefile.someextension
I thought it was silly actually that website urls had all the slashes “around the wrong way” … the http:// helped me learn to get it right.
At the end of the day, it doesn’t matter. Even if your income depends on it, it’s not that big of a deal, you should just do what suits you best. Don’t consider some stupid moral principle. Consider practicalities.
Consider your domain’s cookie policy.
I don’t even advocate naming subdomains after your well known services. yuck. www is implied. website, internet, whatever, if you’re smart enough to figure out that the internet is more than pretty pages, you’re probably smart enough to figure out the rest.
http://ftp.whatever ?? that’s the one I’m most concerned about. Far more important than www. – let’s make ftp more accessible, as it’s easier to navigate without pictures, scripts, css, flash and other … bandwidth-eaters.
mail.something ?? pop3?? smtp?? send?? mta?? mx?? trying to predict what someone will type in to reach a given service is simply too hard to bear. Are you gonna set up your mail servers as aliases for all those subdomains??
There’s nothing to stop you, but it won’t normally be necessary. You can’t win them all. When you think you’ve got all your somain-naming bases covered, something else will come up and ruin your well-made plans.
is it just me or did the filter incorrectly add the http prefix to a part of an ftp url?
http://ftp.example.com
sorry to spam
Keep the www, why is everyone so content with ppl being dumb? they should be told what www means and they should bloody-well remember it! viva LEARNING!
From comment #13:
> What do you plan to do when you want to serve HTTP and FTP from separate machines? You can only give example.com *one* IP address
There are at least two possibilities:
a) Routing the requests on different ports to different machines
b) SRV records (then you don’t need an A record at all). Like MX records which let you use another machine for mail, but not restricted to SMTP. SIP, XMPP and some other services already use it, it is also specified for HTTP and FTP (see ). Too bad it’s not widely implemented yet.
The URL in my previous comment was removed because I put it in chevrons. It should read: “(see http://dns-sd.org/ServiceTypes.html)“.
If ‘www’ (in case it shows then that is without the ‘http://’) is needed to clearly mark a website url that returns a website then why don’t we need it for subdomains? You don’t have ‘www.mail.google.com’! I agree that it does make an excellent visual marker but my website domain is ‘mydomain.co.uk’ plain and simple.
From comment #13:
> What do you plan to do when you want to serve HTTP and FTP from separate machines? You can only give example.com *one* IP address.
Au contraire, only the uninitiated believe this. If you actually look at what’s happening behind the scenes on many major sites, you’ll see that they’re doing fancier things than this to handle heavy loads. For example, try running `traceroute` (`tracert` on Windows) for various sites:
quinn$ traceroute google.com
traceroute: Warning: google.com has multiple addresses; using 64.233.187.99
…
quinn$ traceroute apple.com
traceroute: Warning: apple.com has multiple addresses; using 17.149.160.49
…
quinn$ traceroute microsoft.com
traceroute: Warning: microsoft.com has multiple addresses; using 207.46.197.32
…
These sites generally redirect to www sub-domains partially because the least-informed users will easily recognize that form. However, it has been mentioned that Google has several sub-domains (none of which are hosted on a single machine, mind you) and Apple uses store.apple.com and support.apple.com among others, sometimes with a redirect, other times not.
The choice of using www is arbitrary, and I personally find it annoying. In addition, www can be confusing in situations where sub-domains are actually used to denote hierarchy, such as in universities. For example:
byu.edu
cs.byu.edu
dml.cs.byu.edu
Prefixing each of these with “www” is unnecessary, and in my view, somewhat silly. Honestly, if your user can’t figure out that opening a web browser and typing something like google.com (with or without the protocol) will take them to a web page, we’re in a sad state. Enforcing “www” only cements people’s mistaken assumptions that it’s “just part of the address”.
Quinn,
in your example www would be used to denote the difference between the webserver and the intranet that your uni most likely provides..
if i want to connect to computer softy at cs department at byu school id obviously connect to softy.cs.byu.edu. Its really only when i explicitly want to access the wwwserver on that machine that i would type in http://www.softy.cs.byu.edu
theres obviously a difference and browsers arent the only way to access servers. to say that www is superfluous to me shows lack of understanding.
the “www.mail” arguments made are also quite strange.
if i want to connect to the mailserver i would obviously use “mail.” if i connected through the intranet or through some mail application. However a webmail interface has little need to use “www.mail.” since the webinterface seems better off at the webserver and thus should be accessed through http://www.page.com/mail/ whatever and then in turn connect to the mail server (mail.page.com) to access contents and use the services the mail server provides.
Just because you can send emails in your browser it doesnt mean that it should be the mailserver that handles the webinterface aspect of it. I would rather say that it shouldnt since its a webserver job.
i also dont understand the “subdomain” craze where everything all of a sudden is a subdomain just because it comes after (well, before) the address.
in the example http://www.cassie.cs.bye.edu i would never call my dear cassie a subdomain. Cassie is a reference to the explicit resource named cassie. Neither is www a subdomain since its a reference to the service i want my dear cassie to provide. cs however is a subdomain since it refers to a subgrouping under the bye umbrella. as such changing the dns so that i can access cassie through cassie.bye.edu doesnt make cassie a subdomain either since its still a reference to a resource.
example.com and http://www.example.com means two different things like it or not. by stating www you also specify what youre after which makes the servers treat it in the way you wanted.
To just say screw WWW and let the servers guess what we are after makes little sense. Of course its a neat thing that modern browsers actually do this for us today, but that doesnt mean the right way to do it has no real use.
Im all for smart applications that try to figure out what resource we want, but the underlying standards of structure beneath it all makes perfect sense to keep.
As such applications should be “smart” but not only direct us in the right direction but also tell us exactly what that direction is.
Standards are there for a reason even if the end user doesnt get it.
The definition of a subdomain has remained the same since the DNS was introduced in 1983. See RFC 1034.
A resource is identified by a Uniform Resource Identifier, beginning with a scheme name and colon. “http://www.cassie.cs.byu.edu/” is an HTTP resource, not a WWW resource. “http://cassie.cs.byu.edu/” is also an HTTP resource. “www.cassie.cs.byu.edu” is not a resource at all. See RFC 3986.
I would use the MX record.
I must say that most users do rarly have anything else then world-wide-web on their http protocol. Some do, however, have a script to connect to their own FTP servers on http://ftp.domain.toplevel, in that case, http://www.domain.toplevel is a good idea. In many cases however, the protocol can be switched out, no need for the domain to change.
Pingback: www.www.extra-www.org — omglog
while I disagree with you about the www. being preferred, I am completely agree that both http://www.example.com and example.com should resolve to the intended page, and that it should redirect to one or the other, making the page’s url unique.
I see a lot of people complaining that ftp://example.com and http://example.com would have to be hosted on the same physical machine.
My modem has at least 9 different computers connected to it via ethernet switches and wireless routers. Obviously, all traffic directed to my modem’s IP gets routed to all boxes on the LAN, right? Wrong. There are mechanisms in place in TCP/IP and ethernet protocols to avoid this.
What about serving? The magical solution to the one-IP-many-serving-boxes problem is called NAT. Even with a residential-class modem, I’ve been able to direct access from port A to local IP X, and port B to local IP Y. So I could have http://example.com and ftp://example.com sharing the same WAN IP but being physically different boxes with different LAN IPs. And in fact, that is what I do. And it works.
What you wouldn’t be able to do with NAT is separate http://example.com from ftp://example.com:80, but I think most people will agree using the HTTP port for FTP is silly
Here’s my take on this whole business:
http://www.bigfatdave.com/to_www_or_not_to_www.html
If you’d like to have some fun, try doing this:
http://bigfatdave.com/
or this:
http://bigfatdave.com/blog.php
BTW – IO Error: in case you’re wondering where all your visitors have
gone, they’re all seeing server not found errors, because you’ve got a
DNS error.
Peace,
– Dave
As the General Manager of Search Engine Optimization company, this issue is a real one. Having duplicate content on both http://www.example.com and http://example.com is a horror for the commercial value of the site. You end up with competing sites in the search engines. To maximize ranking you need to choose either http://www.example.com or http://example.com as the target of you link building. Consequently, a webmaster needs to decide, www. or not www.
Given any history of a site, choosing NOT www. is usually a major mistake. If you get natural links, the ones the search engines want you to have, most will be www. today so even for a new site it is a better choice. Next you need to redirect to one version of your site. Choosing both options is just stupid for any commercial site or any site that wants people to visit the site. Of course if you do not want people to visit your site, the site should not exist as it does not serve anyone other than perhaps some stupid algorithm developed by Google to promote adwords and increase barriers to entry to competitors.
I believe in yes-www, but seriously, a gentle reminder?
www misleads to world wide web…
To remind people there are other protocols, shouldn’t it be http://http.example.com/.
I use www since I use subdomains. If I had a site without any subdomains, then I would go no-www. I think it is useful since I have gs1.example.com, forum.example.com, svn.example.com, etc… I just think that http://www.example.com while meshed in there is a lot cleaner than just example.com
Protocol understanding fail.
I prefer the non-www. I don’t like seeing duplicate entries in the SERP with www and non-www links to the same page. I do use a 301 redirect for all www to go to non-www on all my domains. I’m hoping this will help clean up the duplicate entries in the search engines the next time the spiders come to my websites.