Re: [xml-dev] XML Base (RE: [xml-dev] Article: Keeping pace with James Clark)

Tim Bray wrote:
> I always thought the strongest argument for xml:base was that the
> HTML base element was so useful; you could move a whole tree of web
> pages, with relative hyperlinks, around in a nice clean way, and use
> "base" to make sure that relative links were fetched from where you
> wanted them fetched from, not based on on relative processing of
> whatever random URL led someone to the first page.

I have to admit that I'm thoroughly confused by people talking about
moving document trees with relative links AND using BASE to ensure that
links came from a particular location.  You can do one or the other at
any given point, but both simultaneously is a contradition.

> So far I'm not seeing an overwhelming demand for xml:base.  On the
> other hand it's simple, easy to understand, architecturally clean,
> and easy to implement.  The jury may stil be out. 

I question whether XML Base is anywhere near "architecturally clean".
It seems to combine value-by-inheritance headaches with the continuing
nightmare of URI absolutization - not clean in any sense that I can
imagine.

On the other hand, if you know of a convenient scheme-agnostic way to
absolutize URLs in XSLT that takes into account their base URI (whether
from base doc or external entity), xml:base, and whether they needed to
be absolutized in the first place, please let me know.
-- 
Simon St.Laurent
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!
http://simonstl.com