Subject: Re: For Approval: Academic Citing License
From: Johannes Kaiser <jkaiser@geo.unizh.ch>
Date: Mon, 27 Sep 2004 06:50:04 +0200

On 27 Sep 2004, at 0:54, Michael Sparks wrote:

> On Sun, 26 Sep 2004, Johannes Kaiser wrote:
> ...
>> I have written a relatively large (about 5000 lines in C and F90)
>> scientific program and would like to distribute it to the wider
>> community by making it open source. However, acknowledgment is vital
>> for a scientific career these days. Thus the license would have to
>> include a condition that all (scientific) publications produced with
>> the help of my program or parts thereof must cite a specific journal
>> article.
>
> How do you define "(scientific) publications" ? Is it a 
> magazine/journal
> that is considered scientific, any article that uses your program's
> output, or any article that is scientific? How about magazines which
> contain articles which cover science related material? (eg New 
> Scientist
> contains lots of articles which might match, but rarely contains
> citations) How about a scientific article in a scientific magazine 
> aimed
> at 7 year olds?

You've got a point here. I was thinking about those journals which are 
used for communication between scientists of the field. In these, all 
articles have citations. So, it would be Science but not New Scientist. 
Obviously, the definition needs a more careful rephrasing.

> To me the key point is you're restricting use of the program. Whilst 
> (say)
> a GPL program is required by the license (2.c) to display a message to 
> the
> user of that program the GPL under certain circumstances, this doesn't
> cover the *output* of the program, just the derived work - the program 
> -
> itself. What your license states is that users of that program are
> restricted in which locations they can use the *results* of your 
> program
> without citation.

You are right, this is the crucial question. May I ask you to look at 
my answer to Evan's mail?

> Suppose furthermore:
>    * A diagram is produced using a program under this license
>    * This diagram is included in a paper in (say) Nature, and the
>      appropriate citation to your program occurs.
>    * Suppose Reuters pick up the paper and place a summary of the paper
>      on their newswire service, along with a reproduction of the image.
>    * Further suppose that New Scientist (or similar) spot the story, 
> and
>      decide to summarise that story, but include the picture. (Various
>      news sources do have standing arrangements to allow use in this
>      manner base on contractual agreements)
>    * Further suppose that CNN (or similar) spot the story, and
>      decide to summarise that story, but include the picture....
>
> How would New Scientist be aware of the restriction on use of the 
> image?
> (*Would* there be that restriction?) Would CNN hit the same 
> restriction?
> Would Reuters? Who would be responsible for telling New Scientist, CNN 
> &
> Reuters? Nature? Why would Nature know - the restriction is placed on 
> the
> user of the program?

My intention is to require, in this example, a citation in Nature. 
Reuters and the others would be free to use the diagram according to 
Nature's regulations on copyright.

> Who owns the copyright on the image/output? One would assume that the
> person who owns the copyright would be the person running the program. 
> As
> a result one would expect *they* get to decide use, not the person who
> produced the program.

And they would be able to decide the use of all material which they 
have published (citing my article in they original publication). BTW: 
The program produces just the numbers (ASCII file or even Fortran 
variable), not their graphical display.

> I'm not going to answer the questions, I'm asking you them to try to 
> get
> you to think again about what you're asking.

I am trying to think. However the ways of copyrights and licenses are 
new for me.

> In my eyes it looks like a restriction of use, and an attempt to 
> restrict
> to redistribution of works produced using your program. Whilst I've 
> seen
> some proprietary software attempt this (such as cover disk versions of
> compilers), I've never seen any open source (or anything I'd call
> opensource) attempt it.

The wider the works would be distributed the better for me...as long as 
it remains traceable to my program.

> Whilst I understand the sentiment, I would suggest you'd be better off
> *requesting* people cite your journal, rather than try *requiring* it.
> If you're paranoid about citations (I don't know your field),

I have got coworkers who are much more paranoid than I am. And I 
understand them, considering their experiences, e.g. not being 
mentioned in Nature even though they clearly did more of 50% of the 
coding. Not a funny subject.

> and if program
> produces images, you could *perhaps* include a tinyurl.com (or similar)
> url link in the image somewhere discreet. This would be trivial to 
> remove
> is someone wanted to, but I doubt many would.

no graphics, just numbers. In almost all applications, the highest 
level routines of my program would be called by the user's own code. 
So, I have no control over the format of the output.

> After all, as I mention
> above, not all scientific publications contain citations (because it
> depends on the defintion of "scientific publication").
>
> FWIW, I'm not a lawyer, this isn't legal advice, I don't make any
> judgement (and it'd be unqualified if I did) on OSD compatibility, but
> hopefully you find it useful on some level :)
>
> Regards,
>
>
> Michael.
>
>


____________________________________________________
JOHANNES W. KAISER, Dr. rer. nat.
Remote Sensing Laboratories,  University of Zurich
Winterthurerstrasse 190, 8057 Zurich, Switzerland
fon: +41 44 63 55149, fax: +41 44 63 56842
http://www.geo.unizh.ch/~jkaiser