In reality the PC still have the 8 digit system for files like "qwertyui.exe" "mydocu~1.bat", they solve the way to add more characters: "file_from_yesterday.exe", after that, files with spaces "music - song.mp3" by replacing a character when it shows in the Explorer than his real character, the real character of a space is this % Why can't they do the same with web addresses? Why cant a "pepsi world!" be a web address? Do you notice that when an image that has spaces in its name is loaded in a browser the spaces characters changes into % in the address bar? Why cannot the % be shown in the address bar of the browser as a space also translate user input, just the space key be translated into the % character and send over the web to the server, and if the server has it send the page to show in your browser. The space replacer could be another reserved character. when you are writing a phrase web address if you will press the space key, you will see that space in your browser, but before sending the address when you press enter, a filter gets activated that changes this "pepsi world" into "pepsi_world.com" and send it to the server.
-- canoro, Aug 05 2007

Wikipedia: Internationalized domain name http://en.wikipedia...nalized_domain_name
Solves a lot of similar problems in an attempt to accomodate non-ascii characters; doesn't allow space, though. [jutta, Aug 06 2007]

The traditional method would be to keep the existing system, and bolt-on and adopt a shiny front-end that allows the sorts of behaviours and translations you describe, while everything continues to work mechanically as normal in the background.

It's quite funny to think that despite all the hype around operating systems, the web and the whole computer thing in general, that the real goings on, the underlying 3-piece architecture (CPU, Memory and I/O) is pretty the same today as it was since John von Neumann's EDIVAC machine in 1951 (albeit a lot smaller, and with fewer vacuum tubes)

All the progress since then has been on improving the user interface (VDU instead of printer, keyboards and mouse instead of punchcards) - but the nuts and bolts remain almost exactly the same.
-- zen_tom, Aug 05 2007


The same could be said about the car.
-- wagster, Aug 05 2007


//bolt-on and adopt a shiny front-end//

Which only brings up the question of who changes the Internet's oil, and how often?
-- nuclear hobo, Aug 05 2007


I suspect it leaks, as evidenced by the stains found underneath it every morning, and more is put in every week. No need to change it.
-- normzone, Aug 05 2007


I pretty much mirror Ian's sentiments on the subject. Insofar as a small change may affect larger outcomes, how 'bout this: The TLD name parser reads the first and second characters of the address to determine in there is a case discrepancy. 1st character upper case, 2nd character lower case, the parser interprets the address as a phrase.   Serial consistency in the characters is parsed as ambiguous and the 'probable' results are listed by Google rank, or in other words, ranked by importance. You'd then see in your browser a slew of "IM.com Im.com, I m.com, im.com, i m.com, ..." and can chose what you need.
-- reensure, Aug 06 2007


Currently, "pepsi_world" isn't a DNS host name. (As per RFC952, DNS host names can't contain underscores, even though a lot of Windows machine work with names that contain them.) If you want to make an example of translating a phrase into a DNS name, use "pepsi-world".

I don't think this is inevitable or useful - for two reasons.

One, people cannot spell; and even those who can spell can't punctuate, because punctuation is not pronounced and remembered. It's like capitalization. We need fewer of those invisible things to keep track of and get wrong.

Two, just how many different misspellings of the site do you want to make me register? It's a word. Put characters next to one another. There is no white space in halfbakery.

If you're interested in this, go read up on IDN (see link.) IDN fills a real need - but both it and the implementation of this idea have, or would have, the same architecture, and are hitting, or would hit, the same kinds of problems.
-- jutta, Aug 06 2007


In 2020 we will have graphical web url's. Paste a picture of something into the address bar (which will probably be a whole window by then) and it finds it for you.
-- phundug, Aug 07 2007


The real debate in this area at the moment is whether to alow URLs to contain non-Roman characters. To do so would recognise that a good proportion of the world (e.g China) don't use Roman characters as their primary written script, but would also (depending how it is done) either split the internet into islands of content or (like this idea) force domain owners to register another 20 or so variations of their site name.
-- hippo, Aug 08 2007


Isn't this rather similar to AOL keywords?
-- marklar, Aug 08 2007


[BrauBeaton] You can effectively achieve the results you describe by typing your website name into Google (or your preferred Search engine) and then picking out the versions that fit. I've been using Google in this way for ages i.e. not as a search, but as a convenient method of getting to half-remembered website addresses. Even if I do know the address, if I'm on a machine that defaults to the Google search page, I'll enter my website name into Google and click into it from the results, rather than enter the full name directly into the Address bar - I thought everyone did this already.
-- zen_tom, Aug 08 2007


// I thought everyone did this already.//

*raises hand* I do.
-- k_sra, Aug 08 2007


This is just like AOL keywords. I remember when they used to advertise those. not now. I guess people hated going through AOL's portal.
-- twitch, Aug 08 2007



random, halfbakery