Stream: containers

Topic: rachit0412/dataverse-AI-ready


view this post on Zulip Philip Durbin πŸš€ (May 01 2026 at 14:22):

What do we think about https://github.com/rachit0412/dataverse-AI-ready ?

view this post on Zulip Philip Durbin πŸš€ (May 01 2026 at 14:22):

It's described like this: "A complete Docker-based setup for running Dataverse, the open-source research data repository platform developed at Harvard's Institute for Quantitative Social Science."

view this post on Zulip Philip Durbin πŸš€ (May 01 2026 at 14:24):

It says "AI-ready" in the name of the repo but I can't find any AI in there:

Screenshot 2026-05-01 at 10.23.41β€―AM.png

view this post on Zulip Philip Durbin πŸš€ (May 01 2026 at 14:30):

This issue is related, of course: Document competing containerization efforts and how to chooseΒ #10522

view this post on Zulip Philip Durbin πŸš€ (May 01 2026 at 14:32):

I'm not sure what the repo above offers that we don't already describe in our Container Guide at https://guides.dataverse.org :shrugdog:

view this post on Zulip Philip Durbin πŸš€ (May 01 2026 at 14:32):

That said, I invited the author to join the discussion here. Perhaps we can put this on agenda for a future meeting: https://ct.gdcc.io

view this post on Zulip Oliver Bertuch (May 01 2026 at 15:43):

If people are happy with it, I'm fine with it. That said, I wouldn't trust the content one byte.

I don't know the dev behind this. What I see is a lot of text in specific ways of being written, it reeks of AI generated. I don't know to what degree this is just AI slop or actually tested and verified code. I do see it simply forked the upstream compose file?!

I do acknowledge that the dev seems keen on running these things on Windows (.ps1 scripts, small Powershell bits everywhere). But I don't see much content about the scary details involved with that. Again, this reeks of AI slop and maybe not so much of devops experience.

I do welcome contributions from anyone in any form. With one exception: I don't want to deal with AI slop. I use AI every day, but it's filtered through me and given guardrails by me as an experienced devop. Digging through these documents already started to hurt my brain, as I find them not very meaningful.

Any contributions by humans extending, refactoring, or improving the docs, the container images, the pipelines, the codebase: lets have it, much appreciated! But just creating a fork and adding walls of text to it is not a contribution to the community.

BTW: it doesn't even have a license, so technically they are already violating IP law. If it smells like a duck, quacks like a duck, ... :shrugdog:

view this post on Zulip Philip Durbin πŸš€ (May 01 2026 at 15:44):

quack! :duck:


Last updated: May 30 2026 at 09:11 UTC