Stream: coq-community devs & users

Topic: Publications in metadata


view this post on Zulip Karl Palmskog (Sep 12 2020 at 13:03):

Should we have any rules for listing publications in metadata (meta.ml)? I would vote for something like: "a publication either has a DOI (and then a link can be to any website) or it has no DOI and then the link must be to a permanent archive (arXiv, HAL, etc.)"

view this post on Zulip Karl Palmskog (Sep 12 2020 at 13:04):

this would at least give some kind of sanity baseline going forward, so that things don't just disappear

view this post on Zulip Karl Palmskog (Sep 12 2020 at 13:05):

maybe something like this could be enforced in a meta.yml linter

view this post on Zulip Christian Doczkal (Sep 14 2020 at 08:24):

Karl Palmskog said:

"a publication either has a DOI (and then a link can be to any website) or it has no DOI and then the link must be to a permanent archive (arXiv, HAL, etc.)"

  1. Note that HAL does provide DOIs, so a link to HAL without using the DOI would be a little strange.
  2. My own approach for published work has been to use the DOI of the publisher (as a certificate of being published) and let the link point to HAL (to allow easy access).

view this post on Zulip Karl Palmskog (Sep 14 2020 at 08:25):

HAL does not provide DOIs for all publications, example: https://hal.inria.fr/inria-00071689/

and for this one, the DOI indeed goes to the publisher: https://hal.inria.fr/inria-00585203/

view this post on Zulip Karl Palmskog (Sep 14 2020 at 08:28):

so as far as I can see, HAL is like arXiv in this respect, it gives you "stable non-DOI URLs", and the option of inserting a publisher DOI

view this post on Zulip Karl Palmskog (Sep 14 2020 at 08:32):

Christian Doczkal said:

  1. My own approach for published work has been to use the DOI of the publisher (as a certificate of being published) and let the link point to HAL (to allow easy access).

it would be ideal to both have a stable URL to an archive and a DOI to a publisher for everything, but not feasible in my view. This is why I want to allow links to people's websites for preprints as long as they have a DOI.

view this post on Zulip Christian Doczkal (Sep 14 2020 at 08:57):

Karl Palmskog said:

HAL does not provide DOIs for all publications, example: https://hal.inria.fr/inria-00071689/

My bad, I only saw the DOI field without realizing that it was the publishers'.

view this post on Zulip Christian Doczkal (Sep 14 2020 at 09:04):

Karl Palmskog said:

it would be ideal to both have a stable URL to an archive and a DOI to a publisher for everything, but not feasible in my view. This is why I want to allow links to people's websites for preprints as long as they have a DOI.

I wasn't suggesting to make this mandatory. I was more thinking that if we already document guidelines for url/doi, then maybe we say that this is the preferred option (where applicable), assuming we agree that this is indeed the preferred option. After all, there are plenty of other places where one can add URLs to other websites.


Last updated: Feb 05 2023 at 13:02 UTC