URL naming protocol?

July 13, 2005 2:33 PM

TechnoDunce question: why do some URLs start with "http://www.blah" and some just start with "http://blah"?
posted by tristeza to computers & internet (33 comments total)

In traditional naming schemes, "www" is the name of the particular server to which you are containing. Thus, ftp.blah.com might be a different server than www.blah.com or telnet.blah.com or gopher.blah.com. I don't know if this still holds true or not though.
posted by trey at 2:38 PM on July 13, 2005

9 times out of 10 the "www." is not needed.

Sometimes DNS settings require that the "www." be supplied, but I don't know what actually occasions that, and I could be misinformed.
posted by o2b at 2:38 PM on July 13, 2005

www is the name of the server in the blah.com domain. You can setup DNS (the service that matches names to numeric addreses) to have a certain server (www or whatever) answer all requests to a domain: blah.com.
posted by toomuch at 2:38 PM on July 13, 2005

oops i meant *.blah.com in the last sentence
posted by toomuch at 2:39 PM on July 13, 2005

Yeah, it's pretty much a quirk of history. The Internet existed before the WWW, so when people started putting up web servers, they would give those servers a www. name. (Because the whateverdomain.com server may have already existed.) Then, as every domain ended up having a web site, and the web site *was* the domain in many ways, some people started dropping the www. (and/or the http://) to make shorter, more concise advertising.
posted by trevyn at 2:47 PM on July 13, 2005

blah.com and www.blah.com are simply two separate entries in the DNS. They can (and frequently do) point to the same machine, but this is not a requirement.
posted by trevyn at 2:50 PM on July 13, 2005

To the lay-person, simple advise is, it is best to keep the hostname the same when you type it in as it is presented to you. Don't add or remove the www on your own.

As a web developer I've had bugs show up from this. A user added the www on their own, and though the site worked a little, we never anticipated or tested that condition. Cookies got set on the wrong domain scope, and links don't work properly, etc.

Ultimately we (website creators) tend to have to make the www and non www sites both work because people are always adding and removing them on their own initiative, but stop it and save us some work :)

I saw quite a flame war once relating to if its ok to omit the http:// when typing a url into your browser. For that one though, personal preference, I think.
posted by rubin at 3:02 PM on July 13, 2005

The browser assumes you want http if you haven't put a :// at the beginning of the url. So you only need to explicitly specify a scheme if it's not http. As it is, many organizations distinguish their servers by DNS as well as by scheme, so when you type ftp.example.com your browser navigates to ftp://ftp.example.com, so it's even less necessary to specify common schemes.
posted by breath at 3:18 PM on July 13, 2005

Um... let me take a swing at this:

Let's first look at what the parts of a URL are:

"http://silusgrok.blogspot.com/index.html" is the URL to my personal site...

"http://" is the protocol. It declares to any interested party that the rest of what's coming is unencrypted HTTP traffic. It could just as easily be "https://", which is encrypted HTTP traffic, "ftp://", or one of a handful of other — more esoteric protocols.

"silusGROK" is the name of the server — the computer where the site resides — "blogspot" is the domain... and "com" is the top-level domain name. After the slash is the path to the file. In this case, the file has a very short path... but it could just as easily be "folder1/sub-folderX/sub-sub-folder3/index.html".

The "domain" idea is a little weird, but in the early days, an organization would have an entire room full of computers to host their internet material (web site, and other stuff)... and that room would basically be a "domain" of computers. Each computer would then have its own name... which is the server name. And depending on what type of organization that organization was, it would have a specific top-level domain. Of course, things have changed, and most of these distinctions are largely academic... but there you go.

In your case, "www" was a common name for the server which hosted the files used for the "world wide web". But since the most visible thing most servers do these days is serve the web site, most IT guys make sure that if you type a URL without the "www" that you still get the web site. Most, but not all.

Back to the naming... the funny thing in all of this is how strangely it is written:

"protocol://computer.organization.type_of_organization/path_to_file/file.file_type"

So we have:

"protocol://specific.less_specific.even_less_specific/more_specific"

An anally-retentive person, such as myself, would prefer to see their URL look more like this:

"http://com.blogspot.silusGROK/index.html"

But that's not going to happen.
posted by silusGROK at 3:20 PM on July 13, 2005

Oh. One additional point: why do most URLs _still_ have "www" at the front, when they probably don't need them? Well... as a marketing guy, I can tell you that I insist that all URLs are "www.something" so Joe and Jane consumer can easily spot the URL. It's just so easy to spot, that it's worth the extra headache.
posted by silusGROK at 3:23 PM on July 13, 2005

silusGROK might be interested to hear that JANET for one used to do some kind of reversed name scheme. A Google for 'JANET uk.ac' will turn up examples of this.

I'm not sure of the technical details of this though.
posted by edd at 3:29 PM on July 13, 2005

Here is a good argument for keeping the 'www' around.
posted by ubernostrum at 3:35 PM on July 13, 2005

For whatever reason, a lot of servers in Japan do not resolve unless you supply the "www". Try http://yomiuri.co.jp/, a huge newspaper and TV empire.
posted by planetkyoto at 4:06 PM on July 13, 2005

silusGROK: “
An anally-retentive person, such as myself, would prefer to see their URL look more like this:

”http://com.blogspot.silusGROK/index.html“

But that’s not going to happen.”

You mean:

“http://com.blogspot.silusGROK/html.index”

As it’s only one of the various HTML files in that folder, right?
posted by signal at 4:12 PM on July 13, 2005

On the reverse side to the argument that ubernostrum links to, I've seen more than one url get mistyped by a novice user when trying to access a website that's hosted on a machine other than the "default" website for a domain.

Ex. to get to my webmail, I go to "mail.myserver.com". When I gave an account to my mom and told her to go to mail.myserver.com, she typed in "www.mail.myserver.com". I got a call back later that day saying that she didn't know how to get into her e-mail and that something "must be broken".

I believe that most modern web browsers are smart enough to append a "www." onto the beginning of an url if it get's a 404: page not found error back. I don't know if there are any out there that will actually strip off a portion of the address on a 404.

As this question points out, people don't even know what "www" really means, so they blindly follow it and often try to add it in places where it isn't necessary.
posted by freshgroundpepper at 4:16 PM on July 13, 2005

It’s been said previously, but to clarify:

The “www” is just a label pointing to a particular (virtual or real) server within a web domain. Many domains use it to point to their main web server, but this is not a requirement or even an actual convention, just a fairly common practice.
posted by signal at 4:18 PM on July 13, 2005

How does Slashdot force the name to the shortened version when you surf there?
posted by rolypolyman at 4:24 PM on July 13, 2005

To clarify my clarification, re: virtual vs. real servers: “server” refers both to the application that responds to internet information requests and to the physical machine that it runs on, so every server is “real” and “virtual” in some sense. The part before the domain name (e.g.: “www”) points to a specific application running on a specific machine, basically.
posted by signal at 4:26 PM on July 13, 2005

rolypolyman writes "How does Slashdot force the name to the shortened version when you surf there?"

Most likely they use a 301 Redirect
posted by benzo8 at 4:53 PM on July 13, 2005

In my experience, 99 times out of 100 you don't need the "www." Additionally, in 999 times out of 1,000 you don't need the "http://" - why anyone would continue to type all that before the domain amazes me. Why marketing people continue to waste space with "http://www." amazes me even more. Unless you've been in a coma for the last ten years, you know what "microsoft.com" is and how what you're supposed to do with it.

Unfortunately 999 out of 1,000 people don't know that "www" nor "http://" are unnecessary and continue to scractch down on pieces of paper the entire URL with the "http://www." part due to these marketing geniuses that won't let it die.

In short: wriiting "http://www." is pretty much as unnecessary as looking for the keys and TRYING to exactly type "[phone]" on your home phone.
posted by pwb503 at 5:08 PM on July 13, 2005

I agree with the http:// , but the www. is only unnecessary if the domain has been set up that way - it's not a default fall through. Just saying "microsoft.com" is like just saying "1200 Beech Road" without specificying an apartment number - there may be a lot of people living there... And even there's initially only one person living there and 1200 Beech Road works for them, if you want to move someone else in, you get problems, even if they're only ones of consistency.
posted by benzo8 at 5:15 PM on July 13, 2005

I hate to type, virtually never type http://, and seldom type www. Savings= 11 keystrokes, 12 if you count the shift key for the :. I am mildly contemptuous of websites where site.com does not resolve at all. If it resolved to a different service or location, that would make sense, but why anybody would go to the trouble of advertising a site, but not making it as easy as possible for potential customers to get there, is beyond me. My former employer did this. I like that Firefox will autocomplete the .com for me, but I do mostly type it myself, as it's a lot faster.
posted by theora55 at 5:31 PM on July 13, 2005

Pet peeve: the yellow pages website in Chile used to have an ubiquitous jingle on all the radios repeating “amarillas.cl” over and over, but you couldn’t reach the website without adding on “www.”, which the jingle didn’t specify. Guess they couldn’t find a rhyme for “www”.
posted by signal at 6:35 PM on July 13, 2005

Quick note, on all windows machines (and possibly others) using IE 5+ or firefox, typing just ANY straight domain (say, "google" or "metafilter") then pressing and holding 'control' when you push enter will auto-complete the "http://www." and ".com" portions. Similarly, using 'shift' instead will use ".net" instead of ".com"
posted by mystyk at 7:59 PM on July 13, 2005

mystyk: Wow! That was news to me. Thanks!

(and Ctrl-Shift does a ".org", at least on Firefox/Windows)
posted by TimeFactor at 8:32 PM on July 13, 2005

All right, riddle me this: when I enter a full http://www.whatever.com address, but leave out the ":" (http//www.whatever.com), I get Microsoft's site - http://www.microsoft.com/ . And this is in Firefox, so I don't think it's something that's internal to the browser...has Microsoft managed a monopoly grasp on all URL typos of this sort?
posted by attercoppe at 8:35 PM on July 13, 2005

Interestingly, "http//google.com" on Linux using Lynx returns a domain farm called "searchmachine.com", so that's presumably a Windows-ism, not a browser-ism.
posted by arto at 9:14 PM on July 13, 2005

Another quick note to follow up on mystyk's above. You can use "alt-D" in modern browsers (at least firefox and IE, not sure about others) to give focus to the address bar. That makes it really easy when browsing to quickly hit alt-D, type metafilter, and then hit ctrl-enter to get to http://www.metafilter.com lickety-split.
posted by freshgroundpepper at 9:20 PM on July 13, 2005

attercoppe:
I think FF is ignoring "//www.whatever.com" and treats the bare "http" as a search string to Google. (It appears that the default behavior in FF for a word--ie, not a web address--is to feed it to Google and select the first result. For whatever reason, MS's site is the first hit when searching for "http").
posted by MikeKD at 9:23 PM on July 13, 2005

Actualy, I once had a website that would redirect people from http://www.domain.com to http://domain.com. The reason was I basicaly had two web servers which each needed to be configured properly for each virtual host so I basicaly just gave up trying to get them to do the same thing, and put in the redirect. Before that, people who supplied a "www" would get an error message.
posted by delmoi at 10:03 PM on July 13, 2005

Leaving out the ':' implies (to your browser, at least) that you're trying to reach the machine 'http'. So trying 'http//google.com' is asking the machine 'http' for the file 'google.com' in its root directory (multiple consecutive slashes are redundant, i.e. // = /).

If you type in a word in Firefox's address bar and hit enter, it appears to do a google search and follow the first link it finds, a la the "I'm feeling lucky" button. So why does the omission of the ':' send you to Microsoft? Do a google search on "http".
posted by bachelor#3 at 2:06 PM on July 14, 2005

...and only now realizing that MikeKD has already answered attercoppe :\
posted by bachelor#3 at 4:16 PM on July 14, 2005

Well, thanks, guys 'n' gals. We now return you to your origanally scheduled thread, courtesy of tristeza...
posted by attercoppe at 8:43 PM on July 14, 2005