Metadata update without version change

[Sebastian Karcher]

Hi everyone,
Currently metadata updates to a published dataset force a version change. As a curator, I would like a way to avoid this, e.g., I'd like to be able to add links to publications to a dataset's metadata without changing the version number.
It is clear to me that _file_ changes need to update dataset versions, but I don't see the need to force those changes on metadata updates. I don't think this is common practice both for other self-service (e.g. Zenodo) or curate (e.g. ICPSR) repositories. This may already be possible via API(?), but it'd be very nice to be able to do via GUI.
If people disagree about this, would it be possible to make this configurable?

Thanks,
Sebastian

[Pete Meyer]

Hi Sebastian,

The way I understand it, metadata changes trigger a minor version change; file changes trigger a major version change.    I don't strongly agree / strongly disagree - but I suspect it would be easier if it was admin-only instead of curator-only.  Would that make a difference?

I can see cases where this would be useful to a curator; I can also see ways this could go wrong (automated system not seeing a link/identifier in a dataset because it thinks it's got the most recent version).

Best,
Pete

[Sebastian Karcher]

Thanks, Pete

Correct, metadata forces minor version changes, but I still want to be able to avoid that. I think it's going to look odd to a reader/editor if there are all these different versions of a supplementary dataset.

Yes, I'm happy for this to be admin only or even root admin only (the way delete dataset is, say) -- I was using "curator" as a description of the activity, not necessarily the Dataverse role.

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsub...@googlegroups.com.
To post to this group, send email to dataverse-community@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/eb3bb744-fafa-4a4c-a343-6f30cc2f1b1b%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Sebastian Karcher, PhD
www.sebastiankarcher.com

[Philip Durbin]

Not wanting a version bump for a small metadata change reminds me of this:

"I'd like to be able to fix this typo as if it never occurred, but I'm guessing that opens a whole ball of wax as far as preserving all changes, etc." https://groups.google.com/d/msg/dataverse-community/HYMHkrQ0dUI/1qreu8kABAAJ

Of course, we've also heard that people want multiple versions even when datasets are still in draft: https://github.com/IQSS/dataverse/issues/1002

For the record, there's no magic API call that allows you to make changes to a published version without creating a new version. You could always hack on the database (and reindex Solr) but this is not recommended! :)

At least in Dataverse 4 we introduced the concept of a "minor" version bump. You can move from 1.0 to 1.1 to add your links to publications. In DVN 3 you would have had to go from version 1 to version 2.

I doubt I'm talking you out of what you want so please feel free to create to GitHub issue if you are so inclined. You do a good job of writing up these ideas!

Phil

On Mon, Mar 12, 2018 at 7:47 PM, Sebastian Karcher <kar...@u.northwestern.edu> wrote:

Thanks, Pete

Correct, metadata forces minor version changes, but I still want to be able to avoid that. I think it's going to look odd to a reader/editor if there are all these different versions of a supplementary dataset.

Yes, I'm happy for this to be admin only or even root admin only (the way delete dataset is, say) -- I was using "curator" as a description of the activity, not necessarily the Dataverse role.

On Mon, Mar 12, 2018 at 4:27 PM, Pete Meyer <me...@hkl.hms.harvard.edu> wrote:

Hi Sebastian,

The way I understand it, metadata changes trigger a minor version change; file changes trigger a major version change.    I don't strongly agree / strongly disagree - but I suspect it would be easier if it was admin-only instead of curator-only.  Would that make a difference?

I can see cases where this would be useful to a curator; I can also see ways this could go wrong (automated system not seeing a link/identifier in a dataset because it thinks it's got the most recent version).

Best,
Pete

On Saturday, March 10, 2018 at 1:33:39 PM UTC-5, Sebastian Karcher wrote:

Hi everyone,
Currently metadata updates to a published dataset force a version change. As a curator, I would like a way to avoid this, e.g., I'd like to be able to add links to publications to a dataset's metadata without changing the version number.
It is clear to me that _file_ changes need to update dataset versions, but I don't see the need to force those changes on metadata updates. I don't think this is common practice both for other self-service (e.g. Zenodo) or curate (e.g. ICPSR) repositories. This may already be possible via API(?), but it'd be very nice to be able to do via GUI.
If people disagree about this, would it be possible to make this configurable?

Thanks,
Sebastian

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsubscribe...@googlegroups.com.

To post to this group, send email to dataverse-community@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/eb3bb744-fafa-4a4c-a343-6f30cc2f1b1b%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Sebastian Karcher, PhD
www.sebastiankarcher.com

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsub...@googlegroups.com.
To post to this group, send email to dataverse-community@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/CAOSYSD4a2NJtvT5RcJbxLyvMyK6PgZVEd3Y8tv4kTBRzufvi2Q%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--

Philip Durbin
Software Developer for http://dataverse.org
http://www.iq.harvard.edu/people/philip-durbin