# Should we consider grandfathering old R (and other programming-heavy) posts?

I have noticed quite a lot of too-old-to-migrate software-heavy questions (mostly R, but not entirely) coming through the close review queue.

They are often clearly off-topic by current standards and I generally see merit to marking them as such. For example, they can be confusing to new users (and I note they are often highly-visited pages) in terms of what the scope of our site is. Some examples of the type of question I am talking about, though not all closed recently, are here, here, and here.

I have generally voted for closing them. But it has crossed my mind there is a disadvantage to this, which seems to have been the pattern when such votes have occurred: answers to coding questions can get out of date. People who use ggplot2 in R will be well aware that changes to that package have invalidated many of the (irritatingly SEO-friendly) answers on SO — as well as some of the older ones on CV. (In fact, on CV, the issue is in some sense worse, because these tend to be our older inactive questions so are less likely to have been updated or attracted new answers than those on SO.) This is just one example and there is no reason to expect it to be unique. Closed questions, if not deleted, are going to leave old answers to ossify and unable to to be supplemented by fresher answers.

I have my own views about what the scope of this site should include regarding computing questions, but that is a separate matter from what I am raising here, since I am not calling for new questions to follow any different criteria to what we have at the moment. But for old computing questions, from the early days of the site before consensus about our current (somewhat narrower) focus emerged, I wonder whether we should consider being more liberal with letting them stand open — perhaps marked as "grandfathered" in some way so that new users do not mistake them as currently on-topic. (A "grandfather clause" allows pre-existing exceptions to be exempted from newer standards. This seems to have happened to some extent on the "statistics jokes" thread, for instance.)

• Another meta thread on older posts is somewhat related but seems to be addressing a somewhat different issue (unanswered questions vs computing questions) to this one. – Silverfish Jan 12 '16 at 23:18
• I agree that the borderline has moved (though I haven't always been successful at convincing others that it did). I'd lean toward leaving the older questions alone. – Glen_b Jan 12 '16 at 23:31
• What exactly do you mean by "grandfathering" in your title question? Does it imply "being more liberal with letting them stand open" or rather "being strict in closing them"? – amoeba Jan 13 '16 at 10:38
• Sorry @amoeba, I should have made that clearer - it's a legal term. A grandfather clause allows pre-existing exceptions to be exempted from newer standards. In this case, it would mean allowing older questions (that would once have been considered on-topic but would now not be) to be left alone. – Silverfish Jan 13 '16 at 13:36
• Ah, I see, this makes it clearer, thanks. Regarding your worry that closed old questions might have ossified outdated answers without ever getting fresh ones, one should note that it is still possible to leave correcting comments under answers in the closed threads and it is also possible to edit answers in the closed threads (e.g. to make an update or a caveat note). – amoeba Jan 13 '16 at 13:57
• I'm assuming the downvotes indicate disagreement with the idea of grandfathering idea but it would be good if someone could submit an answer to that effect. Personally my tendency has been to vote to close older coding questions in the review queue but not to actively flag them when I encounter one in the wild. I'm unsure whether grandfathering is desirable but I think it is at least worthy of discussion given the current situation is not entirely satisfactory or self-consistent. – Silverfish Jan 13 '16 at 21:49
• I'd suggest copying the explanation of what "grandfathering" means into the question. (The sense here was utterly new to me.) The most worrying aspect to me is not that we have shifted our boundary over time, but the news that code in several older answers just won't work any more. Whenever that's spotted, there should be a comment at least. Also, we need to think about this quantitatively: if there are hundreds of such questions, we should tend to leave them as is. Untidy corners and side rooms are inevitable in a forum like this. – Nick Cox Jan 14 '16 at 12:26
• @Silverfish, I am one of the downvoters b/c I disagree w/ the suggestion. My answer below attempts to explicate my position. (FTR, I greatly appreciate your efforts on the site.) – gung - Reinstate Monica Jan 14 '16 at 15:53
• It is difficult for me to evaluate and support the idea on this post without being able to see some examples about what you think is off-topic and should not be closed (and perhaps having some counter examples, too). – Andre Silva Jan 14 '16 at 19:23
• @gung No problem - I was just wondering whether the downvotes were to express disapproval of the proposal, or whether this is one of those threads that needs "Yes" and "No" answers to weigh up opinion (meta seems to vary in the way these are handled). – Silverfish Jan 15 '16 at 0:28
• Here's another example - clearly off-topic nowadays, but a useful up-to-date answer's just been given. – Scortchi - Reinstate Monica Jan 18 '16 at 13:29
• Thanks for adding the examples and bringing up the discussion (+1). However, I do not think we should grandfather those, because the costs are high. Like you said if we do not close off-topic posts they can mislead new users, I think things are working like they should be. In the examples you cited, the answers are still valid, so what you argued about answers becoming obsolete and misleading do not seem to be a strong case (at least for now). – Andre Silva Jan 18 '16 at 15:25
• @Scortchi That is a particularly good example, thank you. That fresh answer actually refers to a new package so would not have worked well as a comment since it's not just an update to earlier answers (particularly since that would have prevented including an image of the output, which is important in data visualization questions). Had the question been closed earlier, such a helpful addition would not have been possible. – Silverfish Jan 18 '16 at 20:10
• Silverfish and @Scortchi: Moreover, this new answer has by now been accepted by the OP. – amoeba Jan 19 '16 at 23:39