These are key questions.

On Fri, Feb 25, 2011 at 7:38 AM, Steffen Neumann <[email protected]>wrote:

> Hi,
>
> thanks for all the insightful comments, it's a topic I'd love
> to avoid, but I recognize that it needs to be dealt with.
>
> On Thu, 2011-02-24 at 15:07 -0800, richard apodaca wrote:
> ...
> > 1) to pick an open license that will work for the entire site. CC0
> > would be my preference:
> >
> http://depth-first.com/articles/2009/11/10/chempedia-data-downloads-free-as-in-free/
>
> Yes - or PDDL. CC0 is more common.


> On the one hand I really like CC0, because it makes life hassle-free
> (especially in mashing up different DBs) and using it will encourage
> others to do the same.
>

That is the whole benefit.


> In a not-so-ideal world, a company might download the CC0
> and include it into their MS library, and sell it for $$$.
>
> Now, this wouldn't be the worst, but likely their customers
> wouldn't bother to contribute their own spectra to MassBank,
> because the spectra they got were possibly quite expensive
> (especially if they have to buy their own data back!).
>
> The problem is that *licences* don't solve the problem. A licence only says
that IF you take someone to court then there is a document that *MAY* (not
WIIL) allow you to get some form of redress in some jurisdiction.

The F/OSS community accepted the problem of free-riding many years ago and
this is just the same. We have made all our software Libre: JUMBO, OSCAR,
OPSIN, Lensfield, etc.. We know, for example, that many companies use OSCAR
without contributing to the community. That's a slight pity and I regret
that the pharma industry has been so slow to contribute much to the F/OSS
community. There are excellent examples - in the BlueOb - but most are
simply consumers.

But this is the price we willingly pay for FOSS. It means that we can all
use each others's software without worrying wheter we can use if for
teaching (yes, that's a commercial activity), include it in books (yes,
commercial), use in in company-funded projects in academia (yes, that's
commercial). So please DONT use NC. It takes a bit of courage but it is
worth it.

The real benefit is that we create an ecosystem  which is better than the
commercial offerings. We are now doing that. Jmol is used everywhere (who
buys a commercial viewer). Avogradro will do the same. OSCAR is widely used.
The Quixote project intends to demonstrate a radically new approach to
compchem knowledgebases that will give everyone Libre access to compchem.

Let's get these tools into the educational system. Why shouldn't all MS
teaching be based on MassBank? The take up could be quite rapid. We plan the
same for Quixote. When graduates have all used F/OSS systems in their study
they'll want to continue with them.

In short, I don't mind of the MassBank data is bundled
> with instrument software if they keep it free,
> and if the vendor encourages the customers
> to contribute to that community effort.
>

You can only apply moral/social pressure ("community norms") not legal
pressure. When a community accepts the role of Open Data (as bioscience has
done) it becomes mature in the C21.


> > 2) ... 3) ... 4) Internally and externally flag those records openly
> > licensed, and only allow bulk download of openly licensed records.
>
> Yes, that was my train of thought. BTW, this would mean two different
> licenses (free and non-free), and would be halfway to microlicensing,
> wouldn't it ?
>
> It would be impossible to manage


> On Thu, 2011-02-24 at 09:51 -0800, Craig James wrote:
> ...
> > On top of that, you might be able to make the case that by
> > contributing in the first place, people were giving their spectra into
> > the public domain.  Was there any doubt that the library was intended
> > to be widely distributed?
>
> Some of the data was contributed wit the expressed statement
> that they were *not* freely distributable, and we have
> to accept/comply with that.
>

In which case I think you should split the database. It's technically
impossible for users to manage differential licences. The NC-contributors
will see thatr their data are less used and - I hope - can be persuaded to
change their mind.

NC is a nowin-nowin situtation. Only the unscrupulous gain (they won't
honour the licence and you won't be able to police them)


>
> On Thu, 2011-02-24 at 15:07 -0800, richard apodaca wrote:
> ...
> > It's a tricky area, but the data in your collection may not even be
> > covered by copyright laws. No original copyright means no license
> > necessary.
>
> I am also uncertain about spectra that were placed
> as supplementary material to some journal articles:
> am I free to convert data from a journal to MassBank records ?
> Is that related to the question whether *the data* is copyrighted ?
>

The best answer is NO - they are not copyright. I have been through this a
lot. Crystaleye scrapes data from all publishers - has done so for 5 years
and no-one has complained. They have no reason to and they would get huge
negative reaction.

I have taken this up with the STM publishers association and they are clear
on this point. facts cannot be copyrighted.

And a promise. If you make the spectra available in JCAMP we can convert
them to CML. That adds a whole new dimension to being able to link them to
molecules, annotate molecule->spectra and spectra->molecule.

But I can only do it for fully Open Data.

P.



-- 
Peter Murray-Rust
Reader in Molecular Informatics
Unilever Centre, Dep. Of Chemistry
University of Cambridge
CB2 1EW, UK
+44-1223-763069
------------------------------------------------------------------------------
Free Software Download: Index, Search & Analyze Logs and other IT data in 
Real-Time with Splunk. Collect, index and harness all the fast moving IT data 
generated by your applications, servers and devices whether physical, virtual
or in the cloud. Deliver compliance at lower cost and gain new business 
insights. http://p.sf.net/sfu/splunk-dev2dev 
_______________________________________________
Blueobelisk-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/blueobelisk-discuss

Reply via email to