Misunderstood requirements preventing the use of OpenRefine

I receive more and more messages from colleagues working in the GLAM sector, stating that they are not allowed to use OpenRefine for (sometimes) quite the strange reasons.

I will collect some "blockers" here and you may supplement them.

Maybe in the future we will have some FAQ or other document, where we can defuse these "blockers".

No particular order:

  • Entity has no Java License
  • Not allowed to install Java
  • Data is processed on the "internet"
  • It is a web app and therefore accessible from external computers
  • Opens ports on the host
  • No commercial support available
  • Unsafe because source code is openly available
1 Like
  • No painless way to easily backup and restore project files (another from an IT pro that just didn't like having to move a mouse too much, or manipulate ZIP files on his network)
1 Like

@b2m thanks for sharing.
What happens after you address their questions? Are they willing to install OpenRefine?

I’m also curious about the backgrounds of the individuals raising these concerns. Are they the actual users, their managers, or representatives from the IT department of the institution?

We have a FAQ, but it combines various topics. I believe it would make more sense to split its content based on the user journey (pre-install, post-install, etc.).

So the "usual" workflow is like this:

  1. People hear/see/watch OpenRefine and want to use it too.
  2. They ask their IT department for permission to use OpenRefine.
  3. IT department declines with one of the reasons above.
  4. People come back to me to ask for arguments to convince their IT department.

They ask for some "official source" for my claims that their IT department got some things about OpenRefine wrong.

This is interesting; thank you for documenting it. I suggest we create a dedicated FAQ page specifically for IT department questions as part of our documentation. It should serve as a resource that users can easily reference and share with their IT department. This page could also encourage users to reach out to the community on the forum if they have additional questions.

1 Like

I performed steps 1-3 of the workflow mentioned by b2m. My IT dept justified their refusal of OpenRefine with the following: Setting up automatic update of OpenRefine does not seem to be possible. Therefore, offering OpenRefine on the Employer Portal is problematic, for security reasons, and because of need for automation.
Does this make sense, or not?

1 Like

Interesting! OpenRefine is available from various package repositories, including ones for Windows (Chocolatey) and MacOS (Homebrew):

I wonder if your IT department requires it to be published elsewhere for them to benefit from automated updates?

Incidentally, Microsoft has a package manager now with WinGet. It would be nice for OpenRefine to submit release packages to the Windows Package Manager app repository (not the Microsoft Store catalog - a different source repo that can also be pointed to with WinGet)

Steps are here if someone in the community wants to volunteer:

1 Like

This has been my experience also, as a systems librarian in a university.

We’ve progressed past concerns about no vendor support/maintenance, and have gotten to a point where there are two main concerns. Any advice (particularly on the first point) would be much appreciated.

  1. OpenRefine cannot be detected by Microsoft Defender as “OpenRefine uses a non-standard installation architecture and lacks a digital signature for its CPE, which prevents Microsoft Defender—an automated system—from identifying it. Since we have no way to manually intervene, this creates challenges for vulnerability management.”
  2. Manual processes for installing updates (I have referred the IT team to the information on this thread about OpenRefine being available from various package repositories and waiting to see if this helps).

Hi @katherineobrien , welcome to the forum! Regarding the issues with Windows Defender, we have an open issue regarding signing the Windows files here: Digitally sign openrefine.exe to eliminate extra security warnings on Windows · Issue #3003 · OpenRefine/OpenRefine · GitHub. However, this GitHub issue predates some more recent work to improve the install process on Windows. @Martin and I spoke this morning about trying to prioritize this work since this is a recurring issue for our users, but I can’t commit to a timeline for resolving this issue.

A question for Windows users more generally: aside from signing the installer and executable during our release process, would submitting the files for each release resolve this issue? While I think the ideal situation is that we have all of this automated, perhaps this would be a reasonable stopgap measure to make life easier for our Windows constituency.