Just to follow up on what Buddy and Jonathon have said.
Jonathan Guyer wrote:
> I'm inclined to think that producing some general "ThermoPy" interface
> is the right level of granularity. An interface at all is already
> provided by TCS AB and anything FiPy-specific seems too nerrow. It
> should be easier to get buy-in from OOF and other Pythonic projects and
> hopefully from Thermo-Calc themselves if we set out to make a generic
> Python module. I am troubled by this remark in their copyright:
>
> Thermo-Calc Software AB has exclusive rights for further developing
> and marketing all
> kinds of versions of Thermo-Calc and DICTRA
> software/database/interface packages, worldwide.
>
> Whether that's enforceable or not, I think we're better off working with
> them. Any Python interface would still require the user to own TC/TQ and
> Thermo-Calc itself, but if they object to us doing it, or even to it
> being Open Source, I don't think NIST could get involved.
Damm, Edward F. (E. Buddy) wrote:
> Regarding the question of how this would be used.... For my purposes, I
want to use TC to feed the appropriate source terms with the
thermodynamic quantites in fipy, and feed mobility terms to the
diffusion. I'd envision using it only in conjuction with fipy, but if
general access through Python is good for other things too, then that
makes sense too.
> I've asked Paul Mason at ThermoCalc about his thoughts on the copyright
issues. Hopefully Johns instincts are right. A user has to still buy
the TC interface for it to work, and this would then only help with TC
interface sales.
The Thermo-Calc engine, TQ-Interface (library, dll etc) and the
commercial databases are protected by copyright and are license
controlled using the Flexlm license management system. So any user that
wished to use Thermo-Calc or TQ etc in this manner would need a license
for the software. But over and above this, anything that an individual
user codes up, or anything that NIST provides in terms of an additional
interface/library that interacts with TQ would not be something that is
copyrighted or licensed by us.
I may need a better understanding of exactly what is being suggested
here, but my initial reaction is that a ThermoPy type interface which
made calls to TQ (and just saved a Fi-Py user from having to code it all
up themselves) could be open source. It just wouldn't be able to
interact directly with Thermo-Calc without the user having a license for
Thermo-Calc or the TQ-Interface.
With regard to databases (thermodynamic and kinetic), the TQ-Interface
allows Thermo-Calc users to interact with proprietary/commercial
databases and for a user to access these databases, they would need the
appropriate license. However, the Thermo-Calc database structure is
basically an open format and there are data available (e.g. at NIST -
Ursula Kattner/Carrie Campbell) which can be accessed without the need
for a database license and can be freely distributed. This just depends
on what people want to do. In other words, I am saying that the
Thermo-Calc database structure is like an open platform and hence data
that is not subject to copyright can be exchanged openly between groups
(but licensees of copyrighted databases are not allowed to simply write
out the data and share that) - if I need to clarify this, please let me
know and I will do so.
I hope this helps move the discussions forward.
In summary, let us know what we can do to help and we would be
interested in participating in a Workshop if this goes ahead.
Best regards
Paul
--
Paul Mason
President, Thermo-Calc Software Inc
4160 Washington Road, Suite 230
McMurray, PA 15317
Office Tel: (724) 731 0074
Office Fax: (724) 731 0078
Cell: (724) 518 7334
Website: http://www.thermocalc.com
--
