Re: DOM DOM binding considered harmful,discriminates against open source

John Cowan wrote:
> David Brownell wrote:
> > I think you mean "cannot be shipped as part of" ... an open source
> > application should be able to use an implementation of such interfaces,
> > though it can't ship or implement them.
> Whether Java run-time binding is "linking" within the meaning of the GPL is
> a question.  I incline to think it is; if so, no GPLed program can use an
> implementation of the DOM through the standard interfaces (which is tantamount
> to not being able to use it at all).

Where perhaps not all readers are aware that OSD is distinct
from GPL, and GPL has a slightly ... "stronger" stance on some
issues than the OSD, although GPL does conform to OSD.

> > I think that another way to put this is:  The OSD calls for all
> > the relevant Intellectual Property (IP) to be open, not just to
> > provide access to source code.
> This is not the case; in particular, the OSD deliberately says nothing about
> open trademarks, and it is quite common for licenses (even the MIT/BSD)
> to exclude trademarks explicitly. In this case, disallowing the use of "W3C"
> (as in the package name) is clearly allowed.

Point taken -- I was focussing on other aspects of the IP, namely
the interface evolution process.  Which is an IP (copyright in
this case) issue.  You're right that trademarks are a separate
form of IP (as are patents, trade secrets, and surely more).

> > That's a problem for any "standards" organization that wants to
> > control evolution of such IP, since it requires such control be
> > yielded.
> There is no need to yield control over the binding of a particular
> version to the W3C Level 1 DOM, while still allowing modification of the
> Java code itself.  It simply isn't a DOM interface any more, then.

I'm not following you at all when you write that.  The interface
is what the DOM standardizes, and one form of it is expressed as
Java code.

There _is_ a need to yield some control over that; minimally, to let
the source be patched.  Possibly with a mandated name change, with
the "W3C" and/or "DOM" names removed, which clearly forgoes all
binary compatibility with the W3C APIs.  Clause 4.

> >           This is similar to the issue that Sun's "Community
> > Source" licensing has hit -- interface/code evolution isn't in
> > the hands of the community, unapproved forking is prohibited.
> I don't agree that it's particularly similar.  The SCS license is trying to
> protect implementation, whereas what I am talking about is interface.

Both protect interfaces ... SCSL also protects implementation.

Similar in that the control over evolution is a closed process,
not that they're identical.

- dave

Received on Tuesday, 9 November 1999 17:14:02 UTC