[lsb-discuss] LSB conf call notes for 2008-07-30

R P Herrold herrold at owlriver.com
Mon Aug 11 13:49:14 PDT 2008


On Mon, 11 Aug 2008, Theodore Tso wrote:

> On Mon, Aug 11, 2008 at 07:10:04PM +0100, Alan Cox wrote:

>>> Tso earlier asked:
>>> Under the circumstanes where a liticant might sue Sun, 
>>> they can also sue CentOS directly.  If you're not prepared 
>>> for that eventuality,

If a distribution have not undertaken to indemnify (viz., 
under the two licenses from Sun I raised issue with), and no 
other relationship giving rise to liability (say, contractual 
privity) exists, the distribution need not address the issue, 
beyond the 'background noise' that anyone with a filing fee 
may sue anyone.

>> Cox commenting: 
>> Only if CentOS is required by the LSB spec to ship a TCK 
>> verified trademark Java product. At which point the 
>> producers of the LSB may also have some additional 
>> liabilities - so it isn't in their interest either really 
>> is it ?

* nod *  risk identification and control for the LSB, and for 
the 'distributions' is my concern.

> Tso again:
> Even if we don't require a TCK verified Java product, to the 
> extent that Fedora is shipping any kind of Java VM, they are 
> open to that exposure.

* hunh * I do not purport to speak for Fedora.  As you 'cc'd 
me directly, I take it you are asking about only the Trademark 
issue.  Identical behaviours written to a observed or 
published specification, or reverse engineered from a 
clean-room project, or however, were long ago judicially found 
non-actionable (under copyright) in the Leading Edge 'Asa' 
case; don't use the trademark; address (in FOSS, write around) 
any patents which crop up.  [This is fast and rough -- 
citation omitted, and not at issue here anyway]

Putting that to one side, the (LSB mandated, if adopted) 
undertaking of the indemnification clause, and the exposure to 
the NDA issue, each for TCK issues both drop out; so does a 
third-party (non-Sun) claim for shipping a Sun 'vetted' 
product.

CentOS can and already _does_ control the content it rebuilds 
from avaiable sources to a limited subset of what our 
upstream ships in its binary product, based in part on 
carrying a FOSS recognized license.  We exercise much care on 
trademark matters, owing to some unhappy threats in the 
project's early history.

> Heck, if they shippe Kaffe and gcj they are open to a
> certain amount of risk.  And of course, someone could sue a
> distribution if Linux doesn't work correctly, or if a contributor
> sneaks a back door into emacs, or vi, or gcc, or any one of the
> thousands of packages that compose a Linux distribution.


> I ask again --- what makes Java special?

I take this to ask: What is the difference between 'Java' (the 
asserted Sun trademark), v. a java-like set behaviours in a 
specification?

I covered this last week, and Alan raises the issue squarely.

> Cox, again:

> As I keep pointing out if the LSB does not require a TCK 
> verified java(tm) product then the problem goes away for the 
> most part. If Sun say they wont try and go after anyone for 
> /usr/bin/java being a compatible alternative product that 
> goes away too

The difference is that for the first, one needs to proceed 
through a Sun license agreement containing indemnification and 
side clauses asserting no cap on asserted liability when 
'intellectual property' matters are at issue.  Additional 
issues may apply to re-distributing the Sun product as 
outlined.  The second path does not have _those_ pitfalls.

-- Russ herrold



More information about the lsb-discuss mailing list