Frequently Asked Questions
I need help with something that’s not covered here
If you found a bug, have a feature request, or have any other question regarding DCOR, please open an issue in the DCOR repository that in your opinion best matches your query:
If in doubt, please create an issue in the DCOR-help repositry.
Why do I have to specify a creative commons license when uploading data?
This is a choice made by design. For all data that you upload, private or public, you must specify a creative commons license. This ensures that the data can be used by others in the future. DCOR is a free service hosted by an organization serving the public good. If you cannot meet those terms, then you still have the option to host your own DCOR instance (see Self-Hosting).
Can I upload a test dataset somewhere?
For all testing (or development) purposes, you can use the development instance at https://dcor-dev.mpl.mpg.de. All datasets on that server are purged on a regular basis, so feel free to play with it as you see fit.
What happens in the background when I upload a dataset?
For every DC file that you upload, DCOR performs the following tasks in the background:
Generate a condensed version of the original data. This computationally expensive task is necessary to provide fast access to ancillary features, such as volume or principal inertia ratio, to Shape-Out 2 or dclab via the DCOR API. It also allows you to only upload the data you actually recorded (without any disadvantages).
Generate a preview image and extract the configuration for visualization of the data in the web interface.
The original file you uploaded is not changed. You can verify that the uploaded file is identical to the original file on your hard disk by comparing their sha256 sums. The sha256 sum is listed on each resource page under Additional Information.
Please note that, due to this data processing, it may take a few minutes until the preview is visible and the ancillary features are available via the DCOR API.
Why can’t I add resources to existing datasets?
Not being able to modify a finalized dataset is part of the design of DCOR. The idea behind this design choice is that any user who uses a dataset (e.g. for a publication) will always work with the same resources. If you would be able to add resources (or even replace them), then this would impair reproducibility (or at least make things intransparent).
When you upload several resources in a dataset via DCOR-Aid, the DCOR-Aid first creates a draft dataset. When a dataset is a draft, resources may be uploaded and metadata may be edited. After the upload is complete, DCOR-Aid sets the state of the dataset (irreversibly) to active. In the active state, only the following actions are allowed:
setting the visibility of a private dataset to public
changing the license of a dataset to a less restrictive one
Why can’t I delete datasets or resources?
Here, the same arguments about prohibiting the addition of resources to active datasets apply. Scientific data that once have been made available to the public should not be taken down again.
There are exceptions, for instance:
The data are from a blood measurement of a patient at a hospital. The patient never gave consent for his/her data being made publicly available or the patient revoked a corresponding license.
You have uploaded a dataset that contains sensitive patient data that could possibly be used to deanonymize the patient.
In such cases, it is possible to delete entire datasets. However, this procedure requires you to make a clear statement and proviee proof for your claims.