[
Lists Home |
Date Index |
Thread Index
]
On Friday 29 April 2005 23:34, Erik Bruchez wrote:
> Uche Ogbuji wrote:
> >>Has anyone implemented
> >> the full xpointer scheme?
> >
> > We sure the heck haven't. Madness...
>
> Two (naive) questions:
>
> o Why is it madness?
> o Is there an alternative in the works?
I'm not sure what the alternative should be for, but one thing which in my
opinion complicates the xpointer() scheme is the Range extensions it brings
to the otherwise vanilla XPath. The 3rd party scheme xpath1()[1], allowing
selections to be done with plain XPath 1.0, is in my opinion a nice
light-weight alternative to xpointer() which nevertheless is useful. Mozilla
implements it, I was told.
In my opinion, a problem with it is that it does not use a namespace, which
the XPointer Framework requires 3rd party schemes to do. One approach to
solve that, if it is considered a problem, is to lobby for having W3C host
the scheme. I think that is a good idea, because W3C would be standardizing
current practice; embrace a way a specification conflict; and the
specification has sensible content, it wouldn't require much development
except running the W3C machinery..
Merely ventilating the idea.
Cheers,
Frans
PS. There's a newly created XPointer implementation in KDE; it implements
everything except the xpointer() scheme:
http://webcvs.kde.org/kdenonbeta/kdom/xpointer/
1.
http://www.simonstl.com/ietf/draft-stlaurent-xpath-frag-01.html
|
