Code DOI's via Zenodo

[Jean-Paul Pelteret]

Dear all,

I've recently been made aware of Zenodo, a facility to give code and scientific datasets a DOI number. What are your views on giving the tutorials (and, if the individual authors wanted it, the code-gallery examples) a DOI? It could be of benefit (and an "extra reward") to the authors of those particular pieces of work in terms of increasing their citation count. On the other hand, it may go against the spirit of sharing that is a core part of the deal.II project, and many of the tutorials are strongly interlinked so promoting the citation of one tutorial over the host of others that it builds on is not particularly fair.

Best,

J-P

[Timo Heister]

Hey,

I think it is valuable having the option to cite a specific tutorial:
be it as a starting point (if somebody wants to do this) or, to claim
ownership (like I ask my students who contributed a tutorial do). I
already generate zenodo DOIs for deal.II, for example
https://urldefense.proofpoint.com/v2/url?u=https-3A__doi.org_10.5281_zenodo.61430&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=QBOg_BDNEv2oxreDNyMg5qlOXr0VonBbmR-nJnozly4&s=m-02g5Xa9gAtLPafL-gUa_jIap1nABF7b_EYozeVaHA&e=
The problem is that the tutorials are a part of the deal.II repo, so
you can not cite the tutorial directly (and the author list is
incorrect).

I can not think of a good solution for this, so I just create a
citation linking to the .html page of the tutorial.

> --
> You received this message because you are subscribed to the Google Groups
> "deal.II developers" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to dealii-develop...@googlegroups.com.
> For more options, visit https://urldefense.proofpoint.com/v2/url?u=https-3A__groups.google.com_d_optout&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=QBOg_BDNEv2oxreDNyMg5qlOXr0VonBbmR-nJnozly4&s=GI56kIqAf_sSnyPsyorp0kdqttpr8-LTV-1rRhqWlL4&e= .



--
Timo Heister
http://www.math.clemson.edu/~heister/

[Wolfgang Bangerth]

> I think it is valuable having the option to cite a specific tutorial:
> be it as a starting point (if somebody wants to do this) or, to claim
> ownership (like I ask my students who contributed a tutorial do). I
> already generate zenodo DOIs for deal.II, for example
> https://urldefense.proofpoint.com/v2/url?u=https-3A__doi.org_10.5281_zenodo.61430&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=QBOg_BDNEv2oxreDNyMg5qlOXr0VonBbmR-nJnozly4&s=m-02g5Xa9gAtLPafL-gUa_jIap1nABF7b_EYozeVaHA&e=
> The problem is that the tutorials are a part of the deal.II repo, so
> you can not cite the tutorial directly (and the author list is
> incorrect).

But we could create DOIs for the tutorials that really only list the
(original) authors of that program. Presumably, people who cite them would
also cite deal.II itself.

I think a bigger problem is that even if there is a DOI for a tutorial
program, it's still not clear to anyone how exactly to cite it. For example,
the page Timo references above says to use

Wolfgang Bangerth, ., Guido Kanschat, ., Matthias Maier, ., Timo Heister, .,
Martin Kronbichler, ., Luca Heltai, ., … Marco Tezzele, . (2016). dealii:
deal.II version 8.4.2 [Data set]. Zenodo. http://doi.org/10.5281/zenodo.61430

I don't actually know how it came up with the author list -- it may simply be
the complete list of committers to the github repo. It's also not a [Data
set]. So I'm not entirely sure if that would really help people.

Finally, there is of course the question that citations only help if the
citation search engines (mathscinet, google scholar, etc.) actually pick them
up. Is that the case for zenodo?

> I can not think of a good solution for this, so I just create a
> citation linking to the .html page of the tutorial.

That's what I've done as well.

Best
W.

--
------------------------------------------------------------------------
Wolfgang Bangerth email: bang...@colostate.edu
www: http://www.math.colostate.edu/~bangerth/

[Timo Heister]

> But we could create DOIs for the tutorials that really only list the
> (original) authors of that program. Presumably, people who cite them would
> also cite deal.II itself.

I don't know how to do this in a simple way. Zenodo takes a github
repo and a specific release tag and creates a DOI for it. Unless you
manually upload the tutorial source code separately, there is no way
to give it a different title or list of authors.

> I don't actually know how it came up with the author list -- it may simply
> be the complete list of committers to the github repo

Yes, but I can edit this list and remove authors. Any suggestions on
what I should do?

>. It's also not a [Data
> set].

I am not sure how to remove that. It doesn't appear in the bibtex code, though.

[Jean-Paul Pelteret]

Thanks for the replies thus far. Interesting points that you both have raised. I was only introduced to Zenodo this last week and have just created an account, so I wasn't fully aware of how it could / would have to be used.

To put this in context, Andrew and I were having a discussion related to the use of the step-44 tutorial as a basis for further research. We've never had any metrics on how widespread its application has been. For him, without this data, its been difficult to justify as third-party contributors (as we were when we wrote the tutorial) in the writing funding proposals the merits of contributing to open-source software. On a more selfish note, as young researchers it may also have been beneficial to us to boost those annoying but supposedly important citation metrics that we rely on to advance our careers. It goes without saying that, like everyone who contributes, we remain very much in agreement of the altruistic nature of open-source software and the community around it. But on this particular point it would be great if we could somehow inject an extra dose of pragmatism without upsetting the spirit in which the contributions are made in the first place.

 

> But we could create DOIs for the tutorials that really only list the
> (original) authors of that program. Presumably, people who cite them would
> also cite deal.II itself.

Thats pretty much what I envisioned could/would happen. Same goes for the code-gallery contributions, which are more easy to claim ownership over as they're distanced from the main repository. Also, I couldn't see the case for someone citing only one or the other. 

 

I don't know how to do this in a simple way. Zenodo takes a github
repo and a specific release tag and creates a DOI for it. Unless you
manually upload the tutorial source code separately, there is no way
to give it a different title or list of authors.

True, this definitely seems to be the case. You have to upload at least one file when making a new entry in Zenodo.

Thanks for being willing to have this discussion!

Best,

J-P

 

[Wolfgang Bangerth]

> citation metrics that we rely on to advance our careers. It goes without
> saying that, like everyone who contributes, we remain very much in
> agreement of the altruistic nature of open-source software and the
> community around it. But on this particular point it would be great if
> we could somehow inject an extra dose of pragmatism without upsetting
> the spirit in which the contributions are made in the first place.

I think you unnecessarily worry about this. We do want that everyone who
deserves credit gets credit. It's good for all if it's good for one,
because it encourages people to contribute.

The challenge is to find ways to make it happen in practice. I would
definitely support putting a link or note at the top of every tutorial
and code gallery page that explains how to cite the program if we had a
way to create DOIs for each of them. In other words, if you do find a
way to make that happen, I'll write the patch to put in the note!

Cheers

[Timo Heister]

> To put this in context, Andrew and I were having a discussion related to the
> use of the step-44 tutorial as a basis for further research.

You could also consider submitting to https://urldefense.proofpoint.com/v2/url?u=http-3A__www.archnumsoft.org_&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=VLl3-Z5XpHPGLgAFp2LNQI8jgD7xjFq9hP1DOfg5SdE&s=q5Rhh4lreSUkROMqR1828ermzE20AzYNksRY_nvW9t8&e=
Maybe Guido can offer some guidance if this is acceptable even though
it already is a tutorial program. You could always extend it for a
publication there.

> True, this definitely seems to be the case. You have to upload at least one
> file when making a new entry in Zenodo.

You can manually upload the .cc and reference the deal.II repo. Here,
I did an experiment:
https://urldefense.proofpoint.com/v2/url?u=http-3A__doi.org_10.5281_zenodo.400995&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=VLl3-Z5XpHPGLgAFp2LNQI8jgD7xjFq9hP1DOfg5SdE&s=OFCA8kbF8irN9mIMWW_bI38XJuia6TFs1ZcM7UNvvmY&e=

[Jean-Paul Pelteret]

I think you unnecessarily worry about this. 

Ha ha, I don't like thinking that I'm constantly rocking the boat. I'll stop being apologetic now :-)

 

It's good for all if it's good for one, because it encourages people to contribute. 

 👍 Agreed!

You can manually upload the .cc and reference the deal.II repo. Here, 
I did an experiment: 
https://urldefense.proofpoint.com/v2/url?u=http-3A__doi.org_10.5281_zenodo.400995&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=VLl3-Z5XpHPGLgAFp2LNQI8jgD7xjFq9hP1DOfg5SdE&s=OFCA8kbF8irN9mIMWW_bI38XJuia6TFs1ZcM7UNvvmY&e= 

Thats great - pretty much exactly what I was hoping we could do! Just something concise, and that it references to the library makes it unambiguous that its a core part of a bigger project. I would anticipate that we would only do this once per tutorial, since (as best as I can determine) you can't revise the uploaded files. Quick question - what was the rationale behind categorising it as a "data set" as opposed to "software"?

If there are no objections to doing things in this way, then I will volunteer to contact the first authors of the tutorials and code-gallery examples to ask if they want their contributions "captured" in this way. If so, then I would offer to generate the DOI on their behalf should they not wish to create an account and do it themselves. In my opinion, it would be nice to do this before the next release. 

You could also consider submitting to https://urldefense.proofpoint.com/v2/url?u=http-3A__www.archnumsoft.org_&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=VLl3-Z5XpHPGLgAFp2LNQI8jgD7xjFq9hP1DOfg5SdE&s=q5Rhh4lreSUkROMqR1828ermzE20AzYNksRY_nvW9t8&e= 
Maybe Guido can offer some guidance if this is acceptable even though 
it already is a tutorial program. You could always extend it for a 
publication there. 

Thanks very much for the suggestion, Timo. At the time that we wrote it, Andrew and I did consider trying to spin it into a publishable piece of work. We've both gone on to use it as the basis for other work, but like everyone else only cited it by url link (if at all) :-) However, as it stands there is (in my opinion) quite literally nothing novel about it. Its a standard approach formalised many years ago, and its formulation is found in some description in many nonlinear solid mechanics text books and is likely embedded in every commercial FEA code under the sun. It formed the core of the work I was doing at the time, and we just happened to be the first to document it in a deal.II context.

[Wolfgang Bangerth]

> If there are no objections to doing things in this way, then I will volunteer
> to contact the first authors of the tutorials and code-gallery examples to ask
> if they want their contributions "captured" in this way. If so, then I would
> offer to generate the DOI on their behalf should they not wish to create an
> account and do it themselves.

If so, may I suggest a slightly more verbose title of the form
The deal.II tutorial step-56: Geometric multigrid...
rather than the shorter text at
https://zenodo.org/record/400995#.WNEj_CErJhE
?

I think a question that ought to be asked is: Programs are written and
rewritten many times over, by multiple people. There are probably a dozen
people who have contributed to each of steps 1-6, although admittedly the
structure has remained mostly the same. How do you want to reflect the number
of people who have contributed? This is probably not true (to a degree where
it's important to list everyone who has made changes) for many of the later
programs, but it would be worthwhile thinking about it.

And how do you want to reflect if, as Martin recently did with step-37 for
example , someone radically rewrites a program?

> In my opinion, it would be nice to do this
> before the next release.

Oh, Timo and I were thinking of getting the release tasklist started this week
while he's here in Colorado...

Cheers

[Jean-Paul Pelteret]

>> If there are no objections to doing things in this way, then I will volunteer
>> to contact the first authors of the tutorials and code-gallery examples to ask
>> if they want their contributions "captured" in this way. If so, then I would
>> offer to generate the DOI on their behalf should they not wish to create an
>> account and do it themselves.
>
> If so, may I suggest a slightly more verbose title of the form
> The deal.II tutorial step-56: Geometric multigrid...
> rather than the shorter text at
> https://zenodo.org/record/400995#.WNEj_CErJhE
> ?

Ok, I would do this should we choose to move forward with it.

> I think a question that ought to be asked is: Programs are written and rewritten many times over, by multiple people. There are probably a dozen people who have contributed to each of steps 1-6, although admittedly the structure has remained mostly the same. How do you want to reflect the number of people who have contributed?
> This is probably not true (to a degree where it's important to list everyone who has made changes) for many of the later programs, but it would be worthwhile thinking about it.

That is true. If the author list is editable online then this could perhaps be accommodated. Timo, is it possible to edit any of the entries after a request for a DOI has been submitted?

> And how do you want to reflect if, as Martin recently did with step-37 for example , someone radically rewrites a program?

Ok, this would be problematic. Having to upload the code means that “step-XYZ” is in essence frozen and memorialised. A way around it would be to retire each step number that becomes obsolete and is removed from the documentation. I don't see this as a problem since numbering is somewhat immaterial, and as the library has grown has become a little arbitrary. Although the (core) “fundamental" steps are 1-6 (and will likely never change?), the links between the other tutorials are obviously based on timeline of their authorship rather than any strict categorisation.

[Wolfgang Bangerth]

> That is true. If the author list is editable online then this could
> perhaps be accommodated.

An alternative would be to just create a new DOI for a program that has
either undergone significant work by someone not already an author, or
that has been reworked to the point where it is a different program.

Newer releases would then simply reference the new DOI.

Best

[Jean-Paul Pelteret]

An alternative would be to just create a new DOI for a program that has either undergone significant work by someone not already an author, or that has been reworked to the point where it is a different program.

Newer releases would then simply reference the new DOI.

Yes, this would seem to be a sensible solution to the problem.

I’ve started to collate the necessary information for the tutorials and code gallery examples. What I’ve done so far can be found here:

https://docs.google.com/spreadsheets/d/1yaKP8mnVIrVZ6aoGGhi396VGEiY3OR0XfWvgDEeKLio/edit?usp=sharing

Should I create a generic deal.II account on Zenodo for us all to share, and thereby manage the entries related to the tutorials etc? If desirable, we can also create a community page to collect all of the other deal.II related uploads to Zenodo.

[Timo Heister]

Please keep in mind that a zenodo DOI creates a reference to a
specific version of the source code.

> Timo, is it possible to edit any of the entries after a request for a DOI has been submitted?

You can edit most of the meta data including authors. I just updated
the title for example.

> A way around it would be to retire each step number that becomes obsolete and is removed from the documentation.

I would like to avoid producing a large number of unused tutorial numbers.

On Wed, Mar 22, 2017 at 4:18 AM, Jean-Paul Pelteret
<jppel...@gmail.com> wrote:
> An alternative would be to just create a new DOI for a program that has
> either undergone significant work by someone not already an author, or that
> has been reworked to the point where it is a different program.
>
> Newer releases would then simply reference the new DOI.
>
>
> Yes, this would seem to be a sensible solution to the problem.
>
> I’ve started to collate the necessary information for the tutorials and code
> gallery examples. What I’ve done so far can be found here:

> https://urldefense.proofpoint.com/v2/url?u=https-3A__docs.google.com_spreadsheets_d_1yaKP8mnVIrVZ6aoGGhi396VGEiY3OR0XfWvgDEeKLio_edit-3Fusp-3Dsharing&d=DwIFaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=NnEXND6kzagXO4G4yd7vfa9PXN9r53LPFyaNqKp59Wc&s=h02X5YqDxtVMCHz9L3C1ZVE8LYMcgvZtRghLyerUzu8&e=

>
> Should I create a generic deal.II account on Zenodo for us all to share, and
> thereby manage the entries related to the tutorials etc? If desirable, we
> can also create a community page to collect all of the other deal.II related
> uploads to Zenodo.
>

> --
> You received this message because you are subscribed to the Google Groups
> "deal.II developers" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to dealii-develop...@googlegroups.com.

> For more options, visit https://urldefense.proofpoint.com/v2/url?u=https-3A__groups.google.com_d_optout&d=DwIFaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=NnEXND6kzagXO4G4yd7vfa9PXN9r53LPFyaNqKp59Wc&s=11YqQOvmcDPbjEVeG7VkRnJ4wjw3d70yQ7WYBPQ_wFw&e= .

[Timo Heister]

> Quick question - what was the rationale behind categorising it as a "data set" as opposed to "software"?

This seems to be a bug. I selected "software" of course. Note that the
word "data set" doesn't appear in the bibtex text.

On Wed, Mar 22, 2017 at 9:38 AM, Timo Heister <hei...@clemson.edu> wrote:
> Please keep in mind that a zenodo DOI creates a reference to a
> specific version of the source code.
>
>> Timo, is it possible to edit any of the entries after a request for a DOI has been submitted?
>
> You can edit most of the meta data including authors. I just updated
> the title for example.
>
>> A way around it would be to retire each step number that becomes obsolete and is removed from the documentation.
>
> I would like to avoid producing a large number of unused tutorial numbers.
>
>
>
>
> On Wed, Mar 22, 2017 at 4:18 AM, Jean-Paul Pelteret
> <jppel...@gmail.com> wrote:
>> An alternative would be to just create a new DOI for a program that has
>> either undergone significant work by someone not already an author, or that
>> has been reworked to the point where it is a different program.
>>
>> Newer releases would then simply reference the new DOI.
>>
>>
>> Yes, this would seem to be a sensible solution to the problem.
>>
>> I’ve started to collate the necessary information for the tutorials and code
>> gallery examples. What I’ve done so far can be found here:

>> https://urldefense.proofpoint.com/v2/url?u=https-3A__docs.google.com_spreadsheets_d_1yaKP8mnVIrVZ6aoGGhi396VGEiY3OR0XfWvgDEeKLio_edit-3Fusp-3Dsharing&d=DwIFaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=xlkbb1e_zxM1NT74AYTBJog40AxVtOft-lhf-ctjMLY&s=QBiXKDw42wIbaMu8czOhIRC9Jm-aeJhjsZd_WgLmQbg&e=

>>
>> Should I create a generic deal.II account on Zenodo for us all to share, and
>> thereby manage the entries related to the tutorials etc? If desirable, we
>> can also create a community page to collect all of the other deal.II related
>> uploads to Zenodo.
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "deal.II developers" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to dealii-develop...@googlegroups.com.

>> For more options, visit https://urldefense.proofpoint.com/v2/url?u=https-3A__groups.google.com_d_optout&d=DwIFaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=xlkbb1e_zxM1NT74AYTBJog40AxVtOft-lhf-ctjMLY&s=otVUGY9FNvMp6tWhHMf3YfUmE_hlQ_MEfRZg5aM15cY&e= .

[Timo Heister]

JP,

I would like to get back to the question if it is a good idea to
create DOI for all steps:
1. Several tutorials are a community effort and not contributed by a
single individual. This probably holds for 1 - 19. It is difficult to
decide on authorship here. This might be enough of a reason to not
assign DOIs for them.
2. How do you decide between a community and individual effort? Can
the latter become the former at some point?
3. When do you want to create a new DOI? It is weird that you can
change authors for a specfic DOI and I assume google scholar can't
deal with that...
4. What is the threshold to become an author on a tutorial? I hate
applying metrics to this. Can I just go and make some documentation
improvements to become an author on every single tutorial? Or phrased
in a different way: Might assigning authors discourage improving
existing programs?

These are a lot of questions we need to think about.

On Wed, Mar 22, 2017 at 9:46 AM, Timo Heister <hei...@clemson.edu> wrote:
>> Quick question - what was the rationale behind categorising it as a "data set" as opposed to "software"?
>
> This seems to be a bug. I selected "software" of course. Note that the
> word "data set" doesn't appear in the bibtex text.
>
> On Wed, Mar 22, 2017 at 9:38 AM, Timo Heister <hei...@clemson.edu> wrote:
>> Please keep in mind that a zenodo DOI creates a reference to a
>> specific version of the source code.
>>
>>> Timo, is it possible to edit any of the entries after a request for a DOI has been submitted?
>>
>> You can edit most of the meta data including authors. I just updated
>> the title for example.
>>
>>> A way around it would be to retire each step number that becomes obsolete and is removed from the documentation.
>>
>> I would like to avoid producing a large number of unused tutorial numbers.
>>
>>
>>
>>
>> On Wed, Mar 22, 2017 at 4:18 AM, Jean-Paul Pelteret
>> <jppel...@gmail.com> wrote:
>>> An alternative would be to just create a new DOI for a program that has
>>> either undergone significant work by someone not already an author, or that
>>> has been reworked to the point where it is a different program.
>>>
>>> Newer releases would then simply reference the new DOI.
>>>
>>>
>>> Yes, this would seem to be a sensible solution to the problem.
>>>
>>> I’ve started to collate the necessary information for the tutorials and code
>>> gallery examples. What I’ve done so far can be found here:

>>> https://urldefense.proofpoint.com/v2/url?u=https-3A__docs.google.com_spreadsheets_d_1yaKP8mnVIrVZ6aoGGhi396VGEiY3OR0XfWvgDEeKLio_edit-3Fusp-3Dsharing&d=DwIFaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=Z4XPlf-AfJ0AaWG75t_f3G-tMWEAOI10qqR_2aT-KSc&s=6MfQzHpCjq9L8K9YojGf_2AizEbwmk5QWUXL5yNH3pc&e=

>>>
>>> Should I create a generic deal.II account on Zenodo for us all to share, and
>>> thereby manage the entries related to the tutorials etc? If desirable, we
>>> can also create a community page to collect all of the other deal.II related
>>> uploads to Zenodo.
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups
>>> "deal.II developers" group.
>>> To unsubscribe from this group and stop receiving emails from it, send an
>>> email to dealii-develop...@googlegroups.com.

>>> For more options, visit https://urldefense.proofpoint.com/v2/url?u=https-3A__groups.google.com_d_optout&d=DwIFaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=Z4XPlf-AfJ0AaWG75t_f3G-tMWEAOI10qqR_2aT-KSc&s=DivbN68JQR6x5wnjE2H2Y5G7Wc3QydXxEmn0vH4HgRM&e= .

[Jean-Paul Pelteret]

Hi Timo,

Firstly, thanks for all of the comments on the previous questions I raised. They were very helpful.

I acknowledge and agree that every point you’ve listed here requires some (if not extensive) debate. I agree that “formally” assigning authorship to all tutorials would perhaps discourage engagement from the community in improving or maintaining these works, and that is certainly not the point of the exercise. On the other hand, I actively maintain the single tutorial I’ve cowritten (for which I feel I can safely claim “authorship”), and would not want attribution for the minor changes I’ve made to others. But I’ve not spend a considerable amount of time editing these other tutorials, and I wouldn’t know what the threshold for contribution to demand “authorship” would be.

Another way to this issue would be this: There are a number of tutorials that, at the time of submission, were primarily written by non-authors of deal.II (or, for a few special cases, individuals who were not authors at the time of writing :-). I think that steps 15, 21, 24, 25, 28, 29, 30, 33, 35, 41-45, 48, 52, 54, 55, and 57 fall under this category (note: my contribution is included in this list). Some of these already have a tagline at the top of the tutorial like

"This program grew out of a student project by…”

For those tutorials that are not attached to a formal submission to a journal (e.g. step 28), would could offer the authors a DOI and thus some sort of elevated formal recognition for the (significant) initial contribution. This would be included in the header with a notice, such as “The initial version of this tutorial can be cited as DOI…”, and we could even fish out the actual first submission from the project history. Its then clear that subsequent maintenance then (supposedly) falls to the community without further attribution, apart from the typical addition of the contributors name in the maintained source code and the introduction of the tutorial.

Do you think that this might be a more balanced approach, and one which does not infringe too heavily on the community-orientated ideals of deal.II?

Best regards,

J-P

[Timo Heister]

> Do you think that this might be a more balanced approach, and one which does
> not infringe too heavily on the community-orientated ideals of deal.II?

Yes, and I am in favor of this approach. I guess we still have the
choice whether we let the authors themselves create/edit/manage the
publication or we (one of us or a joint account for deal.II) do it
ourselves. I think a single user is the owner of each publication on
zenodo, right?

What is your thought on this? I am happy to create/edit the
publications on zenodo if you want. I also sent a bug report for the
"[data set]" part already.

The second step would be to add a small header to the tutorials as:

The initial version of this tutorial can be cited as

<a href="https://urldefense.proofpoint.com/v2/url?u=https-3A__doi.org_10.5281_zenodo.400995&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=x9TMP2uW1fBowEL52nA00WBGrGAZvLSaAFJtWDjsigM&s=lf15UWgnU61HDyvveWVZ9wZjSbhJpoGVTPVTY8xzs0k&e= "><img
src="https://urldefense.proofpoint.com/v2/url?u=https-3A__zenodo.org_badge_DOI_10.5281_zenodo.400995.svg&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=x9TMP2uW1fBowEL52nA00WBGrGAZvLSaAFJtWDjsigM&s=Wr9jQdTjl-PrRDKgRm10oB0dD52JhpnwDLi2m3xY0VU&e= "
alt="DOI"></a>

[Wolfgang Bangerth]

On 03/29/2017 01:38 PM, Timo Heister wrote:
> The initial version of this tutorial can be cited as
> <a href="https://urldefense.proofpoint.com/v2/url?u=https-3A__doi.org_10.5281_zenodo.400995&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=x9TMP2uW1fBowEL52nA00WBGrGAZvLSaAFJtWDjsigM&s=lf15UWgnU61HDyvveWVZ9wZjSbhJpoGVTPVTY8xzs0k&e=
> "><img
> src="https://urldefense.proofpoint.com/v2/url?u=https-3A__zenodo.org_badge_DOI_10.5281_zenodo.400995.svg&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=x9TMP2uW1fBowEL52nA00WBGrGAZvLSaAFJtWDjsigM&s=Wr9jQdTjl-PrRDKgRm10oB0dD52JhpnwDLi2m3xY0VU&e=
> "
> alt="DOI"></a>

Yes. And this could easily be done using a doxygen macro so that we only
need to list the DOI in the intro.dox file and the text is generated
automatically. Just like the links to the videos.

[Jean-Paul Pelteret]

 Hi Timo and Wolfgang,

On 03/29/2017 01:38 PM, Timo Heister wrote:

Yes, and I am in favor of this approach. I guess we still have the 
choice whether we let the authors themselves create/edit/manage the 
publication or we (one of us or a joint account for deal.II) do it 
ourselves. I think a single user is the owner of each publication on 
zenodo, right? 

 

What is your thought on this?

Sounds perfect. I'd already taken the liberty of creating some DOIs for my contributions as I needed them for a project report, so I think that having the flexibility of the authors creating / maintaining the entries is useful. It does look like there's only one owner per publication. Maybe there is a way to add more but I certainly didn't see any. 

 

I am happy to create/edit the 
publications on zenodo if you want.

If you're willing to then I won't decline the offer. I'll send you the links to those that I've already created for myself.

 

I also sent a bug report for the 
"[data set]" part already. 

Thanks a lot for doing this. Pity that they don't have a solution for it (nor do they notify one of this deficiency).

 

On Wednesday, March 29, 2017 at 9:46:06 PM UTC+2, Wolfgang Bangerth wrote:

Yes. And this could easily be done using a doxygen macro so that we only
need to list the DOI in the intro.dox file and the text is generated
automatically. Just like the links to the videos.

W.

Awesome. That would make this really easy to manage.

[Timo Heister]

I created the zenodo community: https://urldefense.proofpoint.com/v2/url?u=https-3A__zenodo.org_communities_dealii_&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=IXRtTDl1DBR2y7-XpqbXQ6oLYpbtd6vyrW676J-QEfo&s=L2UwGsgVLPIBFJb2twGyzsHZC0IciflHT8hWS8FlhNk&e=

> --
> You received this message because you are subscribed to the Google Groups
> "deal.II developers" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to dealii-develop...@googlegroups.com.

> For more options, visit https://urldefense.proofpoint.com/v2/url?u=https-3A__groups.google.com_d_optout&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=bIZ1PN9GFeiDFVPWRQ7rHwsFOCnh0BavuhRERQ4RDWGqJOxfLu0FfCOxyifWO5wc&m=IXRtTDl1DBR2y7-XpqbXQ6oLYpbtd6vyrW676J-QEfo&s=GuaBOGyhAgyWnl0dV-Wh1G7XfMQhqdc59WFbupF-FCA&e= .

[Jean-Paul Pelteret]

Thanks a lot Timo!