Andrew Gillham writes:

> Just out of curiousity, would "core" accept code from others with
> the same license?  If I were to submit my EISA ethernet card driver,
> which is basically two files, would it go into the tree with my
> copyright and license?  So I would be getting "credit" in advertising,
> etc for a rather simple contribution.  Or, perhaps, someone would
> undertake to re-write my contribution so it became "unencumbered" by
> such a license?

The terms used in the Berkeley license also have the advertising
requirement, so in that sense it's not much different than if you took
the Berkeley license and substituted your own name (which is what a
lot of people do).

The problem here is that the advertising clause was really only
intended for crediting *one* organization.  It was never intended (and
is, honestly, somewhat unwieldly) when there are many copyright
holders using the same license terms.  This is partly why The NetBSD
Foundation exists; we'd like people to assign copyright to TNF so that
we can get back (or at least closer) to only requiring one
attribution.  There are possibly some things we need to work on before
that happens, and developers who are affected by this are welcome to
contact us with suggestions.


But I digress a bit.

The cases you're talking about are a little different, because the
scope of the license terms is different.  E.g., if I'm publishing a
CD-ROM, it's unlikely that the cover is going to mention a specific
Ethernet driver, but it's very likely that it will mention that NetBSD
runs on Alphas.  This is advertising material.  Given the license
terms, even the original Berkeley license terms, this requires me to
include the attribution for Chris, but *not* for you, somewhere on the
cover.  Indeed, it's not clear that any of the vendors of NetBSD,
FreeBSD, or OpenBSD CD-ROMs are actually obeying the letter of all the
license terms, or that they *can*.  This is a real problem, and
something I'd like to get fixed.

(Some people have come up with the vague excuse that it can be in the
booklet or even on the disc, but a consumer isn't going to see this
unless they've already bought the disc, so that simply can't count as
part of the adveertising material.)


FWIW, I objected to Chris's new terms, and went to significant pains
to eliminate even larger problems than Phil suggests might still be
present.  (E.g., it originally required a seperate `This product
includes ...' line for *each* file, which would have essentially
required every consumer of the code to negotiate a waiver.)  But I
don't set the policy by myself.