Here are some impressions of XNS, see http://www.xns.org
I haven't seen their code, but apparantly they have a good,
carefully-designed architecture, and they've implemented it in
Java.
One aspect of XNS which makes it very interesting for DotGNU
that it will be possible to do everything in a strictly
peer-to-peer way. They're planning to provide "root agencies"
which map global names ("XNS names") to URLs, but using those
is entirely optional to using XNS. I'd personally consider this
a main reason that makes me feel comfortable with the idea of
using XNS in DotGNU.
The main problem which has so far prevented release of this code
is that the company which has (so far) funded the development
efforts, and which has been awarded three patents by the USPTO,
needs to find some way to recover their investment and make a
profit from it. It won't work for them to simply give
everything as a gift to the world, but on the other hand they
really want the Free Software and Open Source communities to
embrace XNS. So they need to do something about those patents,
like for example licensing them at zero cost to everyone who
wants to use them in GPL'd code. (BTW I think those patents
could be successfully challenged with Prior Art, but I think
that it would not be right for us to cooperate in any way with
any organisation which uses patents against any kind of Free
Software project. For this reason I think that they need to
freely license their patents in some way before we can start
cooperating with XNSORG in any way.)
The company which has funded development of XNS so far is a
small company, and they will need to justify to their investors
whatever the decisions they make. You can certainly imagine
that this is not an easy decision. It seems to be the main
issue which is so far holding back a release of XNS.
Greetings, Norbert.