bizarre – I just caught myself typing in http://google.maps.com today and thought about this very issue!

It looks like they're doing wildcard subdomain redirects, and using them as search keywords. So, http://blaaargh.base.com/ works too.

If you don't know basic dns, learn it, before calling people names...*cough* [personal attack removed]

yah, things like yahoo.calendar.com work too..

Uhm, Moron Alert, not to burst your bubble but I pointed this out in my post – that mostly any subdomain works with these guys. ("Mostly" because I didn't test *all* words on *all* domains, I just tested with a couple of words.) And yes, I think they're cybersquatters/ domain parkers/ whateveryoumaycall'em.

Well... it (and many phishing attacks, using www.siteitrust.com66.66.66.66) is just a logical consequence of a bizarre decision of reversing the order of _some_ parts of URI. Yes, the protocol is in the beginning. And yes, within the same host, usually directories, subdirectories and files are referenced as usual, from more general to more precise. And the most precise id-anchors are at the end. Even ISO time and date format uses the same convention: YY-MM-DD hh:mm:ss.... Not to mention IP blocks and addresses... or disk:folderfile of DOS hierarchy... or similar paradigms of almost all OSes...

So WHY, omg, it is not tld.domain.host, but host.domain.tld? There is no doubt that there is little hope of reversing this now... but does anybody know why it started being that way?

In email address, the order is also reversed, Veky...

first name – last name @ domain – country
very specific <--> less specific

But while we're unifying the URL protocol, why not also get rid of other meaninglessness?
- "://" is much too confusing – let's use "http/". Even Tim Berner's Lee regrets this notation, he stated!
- why are there sometimes dots (".com") and sometimes slashes ("http : //" or "/folder")?
- the last slash will always be killed
- the order must be reversed from less specific to most specific
- http is a too-complicated abbreviation, so let's use "ht" for hypertext, hts for secure hypertext, ft for file transfers and so on
- the "http"/"ht" is optional, and the browser will start using a different input field for it
- instead of "com/" you can also use "/" as first character

So instead of

http : //calendar.google.com
http : //base.google.com
http : //blog.outer-court.com

... we'll get:

/google/calendar
/google/base
/outer-court/blog

... and so on. Voila!

Wouldn't you then use the chance to finally correct the error you made that you mentioned in http://blogoscoped.com/archive/2005-09-23-n37.html, (5)?

Hmmm... /scoped/blog/google... ;-D

Domain names predated URLs, and are used for purposes other than URLs (e-mail addresses and server hostnames, to name two things). The hierarchical structure where the higher level is to the right has been part of the system from the start (1985), although a different network in the United Kingdom at one point used the opposite order (uk.ac.ox.servername for a server at Oxford, which became servername.ox.ac.uk when they joined the Internet and reversed their address structure accordingly).

Adopting a "simplified" syntax as advocated above would cause lots of problems, such as making it impossible to distinguish absolute from relative URLs, and difficult or impossible to address various things ranging from non-HTTP protocols to addresses in domains not ending in .com (not everything in the world is a commercial entity).