Having links relative to path (i.e. http://domain/path/) - html

Question

Are there commonly accepted ways to have all links and references to images, scripts, stylesheets be relative to some path regardless of current document's URL?
Let's start from the very beginning. I am developing a custom content managing system in PHP. I am using mod_rewrite to redirect all requests like http://domain.com/path/artist/edit/25 to http://domain.com/path/index.php?url=/artist/edit/25. So the part of the URL following http://domain.com/path/ is actually virtual.
I would like all links to be in the format like ... and references to images, scripts, etc. in the format like <link href="ui/css/style.css"...>.
Well, it seems to be possible with:
...
<base href="http://domain.com/path/" />
...
This way I can link to scripts and stylesheets in a way like below:
...
<!-- Custom page style CSS -->
<link href="ui/css/style.css" rel="stylesheet" type='text/css'>
<!-- Support for CSS3 media query in IE8 -->
<script type="text/javascript" src="ui/js/respond.js"></script>
<!-- MooTools 1.6.0 -->
<script type="text/javascript" src="ui/js/MooTools-Core-1.6.0.js"></script>
...
However, AFAIK the <base href=...> should match the current page request (which is http://domain.com/path/artist/edit/25). And it ruins the whole concept.
That's why I need you to clarify:
Is it a commonly accepted practice to have <base href=...> pointing to a directory and not to the current document URL?
Does this practice comply with the requirements for the usage of HTML <base> element?
Will it in any way affect crawlers like Googlebot? Do they require the <base href=...> to match the every particular document URL?
I also would like to know how do you solve the problem of relative links and references to resources when some part of URL is virtual. I have discovered that projects like WordPress tend to completely avoid relative links and go the "absolute links way".

Answer 1

The whole point of the base element is to specify an arbitrary base URL to be used to resolve relative links instead of the current-document URL. Otherwise the element would not make sense since current-document URL is used as the base url by default anyway.
Major crawlers support both absolute and relative URLs as well as the base element. Some shake-and-bake crawlers don’t understand relative URLs and/or don’t support the base element (thus resulting in multiple 404 lines in your server logs, though this is a minor thing).
I would recommend not to use the base element. Relative links tend to be error-prone resulting in wrong resolved URLs while not providing any serious benefits. It’s generally more reasonable and easy to always use absolute URLs.

Answer 2

Is it a commonly accepted practice to have pointing to
a directory and not to the current document URL?
No, it's not common. In fact I'd say it's very uncommon because there are better ways create a logical information architecture of your site without it.
Will it in any way affect crawlers like Googlebot? Do they require the to match the every particular document URL?
It's hard to get the base tag correct and there are ways to do what you want using better methods that are transparent to googlebot etc.
Note, absolute links are what you're seeing in the source but it that does not means that the links physically map to directories and files etc. Using tools like mod_rewrite on apache you can structure your site as many ways as you please with practically any physical filesystem, doing this is also what I'd recommend because as things changes you're not tied to a particular solution. This is also why most php apps send everything through an index.php script, the application then controls the information architecture, not the filesystem.

Answer 3

"base href" can be used without problems, but it is not always the best solution. It is fine if your server will answer requests with diferent server name and paths (e.g. "http://www.example.com/companysection/especificservice" and "http://service.internalnetwork.dev/")
IMHO it's not the best solution for your case.
In the url "http://example.com/path/index.php?url=/artist/edit/25" you want to transform part of the query in a path ( base example.com/path/index.php ?url= )... and this can be a big problem. How are you going to handle querys that also have a query? (receiving a search term or a form GET, for example)
Apache mod_rewrite would be a better option, as Harry answer suggest (or nginx rewrite rules). With it you can easily "transform" a request like http://example.com/path/artist/edit/25?search=something&order=ASC in http://example.com/path/index.php?url=artist/edit/25&search=something&order=ASC
This will give you less problems in the long term.
Check the last example in https://wiki.apache.org/httpd/RewriteQueryString , it's really close to fulfill all your rewriting needs
(you will just need to ensure you handle the rest of query properly)
Take a URL of the form http://example.com/path/var/val and transform
it into a var=val query http://example.com/path?var=val. Essentially
the reverse of the above recipe. This example will work for any valid
three level URL. http://example.com/path/var/val will be transformed
into http://example.com/path?var=val.
RewriteRule ^/path/([^/]+)/([^/]+) /path?$1=$2