[gridengine users] Beware Univa FUD

Dave Love d.love at liverpool.ac.uk
Mon Nov 14 22:18:24 UTC 2011

In case this just looks like a response to defamation, the issues are
important for any FOSS distribution and pointers might be useful to
others.  Anyway, the SoGE distribution is for the community's benefit,
and the community deserves reassurance.  I could expand on this, but
it's long enough and I don't have time or energy.

Fritz Ferstl <fferstl at univa.com> writes:

> Dave,
> you might have wanted to check with me first before you post such messages.
> If you check your exchanges with me from the beginning of this year
> then you will find that I had pointed out to you that not all of what
> you have integrated in the Son of Grid Engine project is kosher.

I don't have a record of such an exchange.  I'd surely recall specific
information about material that wasn't distributable, and the process of
checking, assuming I didn't already know the status of it.

While there were more-or-less nebulous "intellectual property"¹ concerns
expressed on the then steering committee, I relied on reading the
licences, public statements from Sun and Oracle, Univa's actions
presumably on the basis of legal advice, and legal advice provided to
GNU maintainers, for instance.  (Early on I removed items I think
shouldn't have been in the sunsource repo and/or whose licence I
couldn't satisfy <https://arc.liv.ac.uk/trac/SGE/ticket/1282>, but not
at anyone's suggestion.)

> Just because it was accessible from a site offering code under an open
> source license doesn't mean all of its content on that site is under
> that same license.  You still have to look at the fine print.

It sounds as if the then Sun team placed some sort of legal trap while
encouraging people to deal quite freely in the material, but I don't
know what this small print is.  Anyhow their successors haven't sprung
such a trap.  I know the problem with the application of the SISSL, but
it clearly doesn't bother Univa's lawyers or Debian-legal, for instance.

Obviously all the content wasn't under a single licence and there's a
collection of the licences in the distribution.

> In particular you have to check whether things are redistributable and
> may come with copyright issues before you start providing them.

Of course, and if material is under a free licence there aren't any
issues by definition.  For what it's worth, I have been accused of being
"anal" about copyright in connexion with the GNU project by a
maintainer, and rms was annoyed he thought I applied his own legal rules
too strictly.

> The code was open source and under the SISSL and certain binaries were
> under a binary code license that made them redistributable

[But not usable in a useful way, though Sun/Oracle people said the
conditions weren't what the licence said.  Anyhow, those binaries are

> but other pieces, like parts of the documentation for instance, are a
> different matter.

Debian got a licence clarification, particularly about the
documentation, from someone then speaking for Sun
and someone else
In the absence of effort to work on the free wiki docs, the only other
documentation due to Sun that's currently being redistributed
(separately) is part of the Howtos; people can obtain (un-updated)
copies from Univa's distribution if they're worried.  Arguably workshop
proceedings count as documentation, and while I wouldn't have thought
those should be under the sunsource permissive licence, they appear to
be.  I'll obviously abide by authors' wishes if they haven't agreed to
such conditions.

> We've turned every stone before defining the basis for our Univa Grid
> Engine work because we do indemnify our customers and partners

[It's a pity the community turned out not to be the partner we were led
to believe.]

> from
> legal allegations from 3rd parties and we would be liable. I neither
> want users of Grid Engine nor those who provide any version to be in
> trouble so I suggest everybody does some homework.

A lawyer will have to explain how free software users could be in
trouble for receiving a distribution under a free licence.  Anyhow, the
way to save us is to identify allegedly infringing material that has
escaped scrutiny, for which I'll be grateful.  I've no wish to violate
the law of the land -- which I've read -- or use material against
authors' wishes, and I won't stand for accusations of wilfully doing so.

Anyway, let's get this straight.  While doing business in England, Univa
are stating or implying² that I distribute material in (at best) an
irresponsible manner without checking the licences (or whatever
"homework" means),³ and/or that unspecified material infringing (only,
as far as I know) Sun/Oracle copyright is being wilfully distributed
from the University of Liverpool.  This is despite:  at least one Univa
employee had standing to act on it at Oracle, but didn't; Oracle offered
to make it available, apparently for such purposes;⁴ and Univa are
distributing the same (or similar) material under the same licence as
far as recipients can tell.  Right?

The OGS people and distributions like Debian will doubtless want to know
whether they face similar allegations.  If not, why not?

As Univa have this covered, here's a request to supply the source of the
mysql-connector-java-3.0.9-stable-bin.jar binary which they have
distributed to me under its GPL 2 licence.  (I had to remove it
destructively from version control because I couldn't obtain the source


¹ http://www.gnu.org/philosophy/words-to-avoid.html#IntellectualProperty

² With roughly the same effect in English law, as I understand it

³ Despite me discussing licensing details with them

⁴ Around http://gridengine.markmail.org/thread/cag3v4gdckl5eyva

More information about the users mailing list