🏠 back to Observable

Attribution for attached files

I just ran into a minor problem with regard to attached files:

  1. I had attached an image to a notebook and wanted to add attribution.
  2. I couldn’t add the attribution directly to the file because there are no fields for metadata.
  3. On the other hand, adding the information directly to the notebook seemed like a bad solution because I have no way to reliably reference the image: The file alias is generic, and the file can be replaced. As soon as someone forking the notebook replaces the file any attribution becomes stale.

I’d love to hear if others think of this as a problem, and what methods people currently use for attribution of FAs. Some counter arguments to get started:

  • I haven’t worked much with file attachments yet, so I might misjudge how rarely such a situation would actually arise.
  • If someone forks a notebook and replaces files, it’s their responsibility to ensure all attribution is correct.

Thanks for raising this topic.

I feel this issue is more general than file attachments; it also applies to code snippets, for example. My personal recommendation is to include visible attribution in your notebook, and expect others to update the attribution if they change the file (or encourage them to do so by way of a suggestion).

That said, if maintaining attribution becomes too tedious this way, I could maybe see us offering some sort of editable file metadata in the future. But nothing is on the immediate roadmap.