[ntp:questions] Re: Why ntpd is losing out to openntp at BSD

David L. Mills mills at udel.edu
Thu Oct 14 23:34:51 UTC 2004


Thanks for the careful analysis and reply. I copied the statement almost 
verbatim from the MIT copyright statement for Kerberos. As I said, my 
main goal is to preserve the connection with UDel as well as incorporate 
the copyright statemenst of other authors you find in various files. As 
for "losing to openntp," that doesn't bother me a bit. I do, however, 
strongly object to calling it that; it is in brutal fact opensntp. I 
would assume the OpenBSD folks to claim conformance to one specification 
or another, for otherwise it surely is not maintainable.


David Woolley wrote:
> In article <ckkk9f$phc$1 at dewey.udel.edu>, David L. Mills <mills at udel.edu> wrote:
>>The most important issue is that the University of Delaware remains the 
>>copyright holder of this work 
> That is not the issue that they are raising.
>>                              with individual authors acknowledged as 
> Whilst I'm not a lawyer, to me the licence doesn't require other authors
> to surrender copyright.  There is no requirement for contributors to
> assign copyright to the University of Delaware.  Unless you have had
> contributors of non-trivial changes execute an assignment document
> (the Free Software Foundation actually requires a nominal payment, to
> form a valid contract, when doing this), I would say contributors own
> the copyrights in their non-trivial contributions.
> Requiring copyright assignment before inclusion in the official version
> would probably be a good idea, as it makes licence changes and licence
> enforcement easier.  (Making it a requirement on all versions would
> definitely take the software out of the definition of "open source".)
>>listed. If somebody wants to rip it off in other products or sell it on 
>>a CD, 
> Open BSD's lawyer's argument is that the current licence does not permit
> this because the phrase "without fee" can be seen as qualifying "and
> distribute".  CD's are typically distributed for a fee.  They continue
> to distribute the Delaware version from their web site, because no fee
> is charged, but exclude it from their CDs, for which a charge is made.
> (A number of "freeware" licences explicitly ban supply on paid for CDs,
> so it is not unreasonable for their lawyer to assume that this is what
> you meant.)
>>Just be sure the UDel copyright notice is posted somewhere under the hood.
> They wouldn't have a problem with this; it's basically what they
> require of code for which they do own the copyright.  In Europe, you
> would have a moral right to have such a notice, and even in the USA it
> is an open question as to whether it is possible to abandon copyright.
> Generally in the sort of open source circles in which BSD operate, not
> having a copyright notice would be a problem, as it makes auditting the
> copyright licensing difficult.
> The consequence of this dispute about the meaning of "without fee" is
> that a significant segment of the open source community is ganging up
> against ntpd in favour of a poor alternative, which is really SNTP, not
> NTP.  This is being fueled by a certain feeling of righteousness, about
> being truly open source, but also about Open BSD as well.  Open BSD users
> see it as being a very secure system, and are putting that attribute onto
> openntp (partly because it is small, because it isn't a full NTP 
> implementation).
> The non-legal, non-security, argument that they use is that most people
> don't need better than about half a second accuracy (even if the starter
> of this thread is complaining about 10ms errors).  If you add on the
> demands for sub-second initial time setting, step free response to major
> input steps, and cooperating isolated islands (which account for most
> of the current FAQs), the official ntpd is at risk of losing out to
> poorer substitutes.

More information about the questions mailing list