The "off-topic" message when a question is closed currently reads as follows:
This question appears to be off-topic because it focuses onprogramming, debugging, or performing routine operations, or it asksabout obtaining datasets. You could try the support links we maintainor the Open Data site instead.
Should the Open Data site be so prominent in this message?
In my experience on Cross Validated, off-topic questions about programming or software recommendations are (currently at least) much more common than off-topic questions about open datasets. So why not mentioning directly stackoverflow and softwarerecs too in this message?
In addition, it's been my experience that there are currently very few active experienced users on opendata.SE, which has an impact on the website moderation. So I'm wondering if we should be more careful about suggesting asking questions on opendata.SE, when they could also be off-topic there.
As a recent example, there is this question that has been asked today on opendata.SE right after being closed on CrossValidated. I don't know what the community at opendata.SE think about it, but I believe the question might be off-topic there too, and could have been a better fit for the software recommendations website.
So, is there some room for improvement of the CrossValidated "off-topic" message?
Per the comment of Sycorax about the difficulty of adding content to the message, an alternative would be to remove any mention of specific websites, and have a message such as:
This question appears to be off-topic because it focuses onprogramming, debugging, or performing routine operations, or it asksabout obtaining datasets. You could try the support links we maintain instead.
Edit, August 2025:
Another example of someone incorrectly directed to the open data website due to the over-emphasis put on it in the closing message: How to get separate calibration lines for different groups in a model? was closed as being off-topic, which led the asker to ask it on the Open Data website after seeing the closing message https://opendata.stackexchange.com/questions/22035/how-to-get-separate-calibration-lines-for-different-groups-in-a-model .
I left a comment under his question to direct him to stackoverflow instead, but long-term I don't think this is an optimal way to deal with the issue. I think it would be a good thing to eventually change the message to the one suggested by @SextusEmpirus in his answer.