Need scheme relative url clarification - http

Question

I've been reading about url's. Absolute, scheme relative, root relative, location relative.
I still don't understand difference between these two:
//domain.com/index.html - scheme relative
domain.com/index.html - ?
.
Question 1:
Correct me if I am wrong //domain.com/index.html will resolve to absolute url like this:
http://domain.com/index.html
https://domain.com/index.html
ftp://domain.com/index.html
file://domain.com/index.html -- if in email
And browsers will act differently: ie6 doesn't support, ie7,8 will fetch data twice(http https).
.
Question 2:
How will domain.com/index.html resolve? Same as scheme relative url in Q1? Or is it something else?
.
Question 3:
Is there any difference between these url's, what is it and why?
//www.domain.com/index.html
www.domain.com/index.html
.
Question 4:
How will //www.domain.com/index.html resolve?
.
Question 5:
How will www.domain.com/index.html resolve?

Answer 1

It's very easy, looking at URLs like these, to apply your human knowledge of what they probably mean, rather than the much simpler rules implemented by software like web browsers.
The simplest type of URL (or more accurately URI, since some schemes don't represent a Location, only an Identifier) is absolute; it starts with a scheme, then a colon, and no context is needed to resolve it. Examples:
http://example.com
https://www.example.com/foo/bar.baz
http://127.0.0.1:8001
mailto:someone#example.com
data:text/plain,test
urn:example
Then there are location-relative URLs; that is, anything without a scheme, and without a leading slash. These replace everything after the slash in the current context, but leave the rest in place. If the current context is http://example.com/foo/bar.baz, you could have relative URLs like so:
bob.baz -> http://example.com/foo/bob.baz
thing/widget.gizmo -> http://example.com/foo/thing/widget.gizmo
example.com/page -> http://example.com/foo/example.com/page
Note that that last example looks like a domain name at first glance, but is actually exactly the same as all the other relative URLs.
Root-relative URLs, with a leading slash, are similar, but instead of deleting after the last slash, they delete after the first. Given the same context, the previous examples become:
/bob.baz -> http://example.com/bob.baz
/thing/widget.gizmo -> http://example.com/thing/widget.gizmo
/example.com/page -> http://example.com/example.com/page
A root-relative URL could also contain a colon, because the leading slash cannot be part of a scheme prefix:
/foo:bar -> http://example.com/foo:bar
/urn:example -> http://example.com/urn:example
Finally, there are scheme-relative URLs, with two leading slashes. They replace everything after the original double-slash, so keep only the scheme:
if the context is http://example.com/foo/bar then //example.org/bob means http://example.org/bob
if the context is https://example.com/foo/bar then //example.org/bob means https://example.org/bob
if the context is http://example.com, then //foo.bar means http://foo.bar
Note that that last example doesn't look like a domain name to us, but it still follows the same rules. Whether a URL is actually useful is not taken into account when parsing any of the relative forms.
Conventions like "begins with www." and "ends with .com" cannot be relied on, and are not used to determine if a URL is relative or not, so all you need do to answer all your questions is follow this simple set of rules:
If there are two leading slashes, it is scheme relative
If there is one leading slash, it is root relative
If there is no leading slash, but there is a colon, assume it is an absolute URI
If there is no leading slash, and no colon, it is location relative

Answer 2

They are very different. The second one is a relative reference to a path "domain.com/index.html".
WRT "domain.com" vs "www.domain.com": these are simply different host names (or path names in the second variant)