Hello Martin
Thank you for your reply drawing attention to copyright issues, and I
agree with your comments.
I see the ACDD page we use states at the end: "Content is available
under GNU Free Documentation License 1.2
<http://www.gnu.org/copyleft/fdl.html>."
So it seems we must try to re-word everything to not be a copy or, copy
and follow the licence and acknowledge and reproduce it etc. It would
good if we could get a specific release somehow from these obligations,
as all we are trying to do is to implement the very ideas that have been
specified in ACDD. However, the licence does legally cover their
material because they have specified that it does.
In the preamble of the Javadoc we do refer to the ACDD. We could also
add something more specific about the licence that may satisfy it, but
it does get complicated. I have no experience of having to honour a GNU
Free Documentation Licence with material embedded within modifications.
The solution may even be a simple one solved by an introductory
statement of acknowledgement in the licence's terms.
Does Apache have some legal division who could advise us as to best
practice in the circumstances? Other users of such licensed standards
must have come across this and similar problems as well.
Kind regards
Rob
On 30/05/2017 14:54, Martin Desruisseaux wrote:
Hello Rob
Thanks for your observation. Indeed, the javadoc of AttributeNames class
merely repeat documentation that exists elsewhere (ACDD and
CF-Conventions). The ACDD and CF-Convention are the authoritative
sources, so we should comply with whatever they said. The only reason
why we did not reproduced the documentation more extensively is that I
was not sure if it would be compliant with ACDD / CF-Convention
licensing. I presumed that copying a summary sentence could be
considered "fair use" but was not sure if more extensive copying would
still "fair use" under copyright laws.
However given that the ACDD and CF-Conventions authors are mostly public
or non-profit organisations, there is good chances that the copy
conditions are very permissive. It just needs to be verified.
Regards,
Martin
Le 30/05/2017 à 14:19, Rob Wallace a écrit :
Hello
Comparing 'Field Detail' in the Javadoc for the AttributeNames class
with ACDD 1-3, in many cases, the Javadoc merely repeats the same
description as used for its 'Field Summary'. In the ACDD, the field
description is more fully described. See field 'COMMENT' for example:
ACDD under Recommended - "Miscellaneous information about the data,
not captured elsewhere. This attribute is defined in the CF
Conventions (http://cfconventions.org)."
Javadoc - "The "comment" attribute name for miscellaneous information
about the data (Recommended)."
I feel that semantically the two are similar in a general sense only,
but not identical in a fuller sense of the meaning. If not identical
semantically, it means that the ACDD document will always have to be
referred to in order to give a fuller definition and guidance. So then
the Javadoc is not properly the full story.
I feel it is a pity to not include what could be very useful detail in
the 'Field Detail' of the Javadoc. The Apache SIS is software used
for scientific purposes and terms need tight definitions that improve
people's understanding, and Javadoc is to where we all refer when
trying to be clear about our facts. It also reduces arguments and
misunderstanding about the detail - so important when multinational
projects are involved.
What is the development team's policy on this matter?
Kind regards
Rob Wallace
