incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Murray Altheim <>
Subject Re: XNode 1.1 API submission
Date Thu, 18 Sep 2003 23:12:03 GMT
Cliff Schmidt wrote:
> Yes - still just collecting notes and links on the Wiki; I'm sure some
> actual sentences would be more helpful -- sorry that's not ready yet.
> Here's my $0.02 on two other issues (one cent each):
> - In reading the email you wrote to Brian, I felt like you were saying
> that there probably weren't any IP entanglements, but that it was under
> the radar enough not to matter anyway.  Personally, I think you should
> do whatever is necessary to ensure Sun has no claims at all.  The thing
> with Apache is that we want to make sure that when some other 
> organization wants to adopt an ASF product, they should feel confident
> that everything they get strictly complies with the Apache Software 
> License.  "Under the radar" isn't enough to make most organizations
> continue to trust the Apache brand.

First off, thanks everyone for spending the effort -- I didn't want to give
the impression I was banging on the door -- I felt I'd walked into a room
where there was an busy conversation and just needed to figure out where
the men's room was. I'm honestly not one of those people who expect
lightning responses to emails (Jon Bosak told me we all needed to take
back our lives from the 24hr email response-mongering, and I agree).

So anyway, Cliff and everyone, thanks for your suggestions. I don't think
there's much of anything to do as regards Sun, or that could be done. I'm
not even sure how it would be possible to evaluate legally *how* different
the XNode 1.1 API has to be from XNode 1.0 for there to not be IPR issues.

One thing that might be confusing this is that *I* had the intent of
submitting the XNode 1.0 API to ASF while I was at Sun. This intent
was not broadcast outside of the few programmers I was working with,
and was never acted upon by anyone at Sun, including me. Legal never
heard of it, nor were there any requests for forms, etc. The code is
embedded deeply within the Java Registry Server; the only way you'd
ever know it was there was to look at the java class names of the
un-jarred file. There's no public docs on XNode 1.0 from Sun -- it was
in effect the way *I* solved the problem of adding metadata to the
Xindice database nodes we were storing. That's why I called it "under
the radar." People solve problems all the time, and it's always a
question of what constitutes a solution as being a "thing" worthy
of IPR consideration. A class? A specific method?

Apart from the name "XNode", there are substantive differences between
1.0 and 1.1, and I can certainly change the name (there's about 2,400 hits
on Google on "xnode", so I felt that was the least of my worries).
XNode 1.1 shares the idea of a SOAP-like wrapper around an XML document
as a container for metadata. The idea that you'd pass these around
as "nodes" comes from Xindice. The wrapper syntax is similar to SOAP, the
node idea is similar to ideas in Lee Iverson's NODAL project. But these
ideas are not patented, and while the 1.1 API necessarily shares many of
the same methods as 1.0 (e.g., you want get and set methods on creation
and modification timestamps, etc.), they are substantively different APIs.
They don't use the same code, the XNode 1.1 documentation is not taken
from 1.0, etc.  XNode 1.1 has a listener interface that isn't in 1.0.
But they are similar because they both perform basically the same
function of adding a metadata wrapper to an XML document, and were both
written by me, and I use a particular style (as does anyone). I didn't
go through the code and simply change the names. In fact, some of the
class names are the same, but they don't share any code between versions.

In law, as you probably know, there's no real resolution in these
kind of situations absent a judgement from a court. I don't believe
that Sun can claim XNode 1.1 is theirs. As you probably know, there's
probably *some* legal hook by which all the work I've done on my Ph.D.
can be claimed to be Sun's IP, perhaps up until my death. I can't
live that way, and if nobody will take this API I'll publish it
myself and move on. (I don't say that to be snotty, just as a
matter of lifestyle:  I won't live in fear of lawyers. They may come
after you whether there's a case or not, as we've seen in many high
profile cases, like Clinton, or the whole current situation with
web plugins.)  As you might guess, I'm not happy about this whole
situation, which is both trivial and confusing, and probably a bunch
of wind in the end. But I also understand where ASF is coming from.

I just don't know how to proceed further.

 > - You mentioned in a prior post that XNode may have broader application
 > than Xindice, and therefore may be a TLP.  You missed the step in
 > between: being a separate subproject of the XML project.  If this
 > sounds reasonable to you, I would post to the general list at

I'm happy to do so, but I felt I should probably air the issues
surrounding its submission prior to going to any specific project
team (which is I believe what Brian thought too in suggesting I
bring it here rather than continue to with the Xindice team). Oh, and
yes I do think submitting it to the XML project is a reasonable idea.

I'm also planning at some point to open source my Ph.D. project "Ceryle",
which includes an implementation of XNode running over Xindice. The
code is somewhat tied into the Ceryle framework, but could be extricated
into a base that could be reused in other apps. And there is no issue
of IPR (at all) on the implementation code. A complete rewrite from the
beginning, and quite different in a number of areas (better!).


Murray Altheim               
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK                    .

           "The current and future international political
            environment severely constrains this country's
            ability to conduct long-range strike missions." -- DARPA

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message