Re: Absolute BASE

Russell Steven Shawn O'Connor (roconnor@wronski.math.uwaterloo.ca)
Thu, 26 Feb 1998 20:00:53 -0500 (EST)


Date: Thu, 26 Feb 1998 20:00:53 -0500 (EST)
From: "Russell Steven Shawn O'Connor" <roconnor@wronski.math.uwaterloo.ca>
To: www-html@w3.org
Message-ID: <Pine.SOL.3.95q.980226200012.22928A-100000@wronski.math.uwaterloo.ca>
Subject: Re: Absolute BASE

> >Any paticular reason for requiring the HREF attribute in the BASE element
> >to be an absolute URI?  Why not resolve a relative URI according to the
> >document's URI?
> 
> That would be contrary to the purpose of BASE though, wouldn't it? BASE
> DEFINES what the relative URIs are based on, and make the href for
> *base* relative as well would pretty much defeat the purpose. 

I imagine most of the time authors would make the BASE absoulte.  One
case where I could use a relative base is the following.

My home computer is running a HTTP server (jigsaw :-) I copied one file
from one directory to another, but I want all the links to still work, so
I'd normally add a BASE element.  The problem is that internally my
computer has one name (eg rosa.vtl.org), and externally it has another (eg
ppp13.granite.mb.ca) so I can't give an absoulute URI for my base element,
becuase there are two possilbe machine names, not to mention the ppp name
isn't static. Ideally I'd like to add <BASE HREF="/original_directory/">

There may be other uses for a relative BASE as well.  I can't think of
them all.  Perphas the uses may be rare, but the point is that restricting
the HREF attribute in BASE to an absoulte URI gains nothing and loses a
bit.  So it would be better if this restriction is removed in the future. 

-- 
Russell O'Connor                           roconnor@uwaterloo.ca
    <URL:http://www.undergrad.math.uwaterloo.ca/%7Eroconnor/>
"And truth irreversibly destroys the meaning of its own message"
-- Anindita Dutta, "The Paradox of Truth, the Truth of Entropy"