# Increase the complexity of a question once a good answer has been provided for a simplified formulation

I'm wondering if it is considered as good practice to increase the complexity of a question once a good answer has been provided for the original problem. Specifically, assume that in the first place a simplified version of a given problem is presented. In case a satisfactory answer has been provided and the simplified problem is solved, is it possible to edit the original post so as to tackle the actual problem formulation?

Please notice that my situation is not exactly the same as in this question, since I would like to increase the complexity of the same problem, rather than asking an additional question about corollaries of the problem.

EDIT: thank you for your answers. When I say "altering the original question" I mean appending new content, something like "let us now alter the function $f(\cdot)$ so that...". Although I'm trying to solve the problem by myself, I was curious to know what to do when some additional help is needed.

Thanks.

• No it is not.  – Did May 29 '13 at 8:25
• Certainly, the proper course depends on how substantial the alterations to your $f$ are... – J. M. ain't a mathematician May 29 '13 at 8:40
• From $f(\mathcal{J}) = E\left(| \sum_{j \in \mathcal{J}} y^j |\right)$ to $f(\mathcal{J}) = E\left(| \sum_{j \in \mathcal{J}} logit(p^j) y^j |\right)$. Moreover, I'm quite confident that the arguments in the best answer are still valid, apart from the final step. Given that the function $f(\cdot)$ would be slightly modified and the provided proof is substantially unaltered apart from the final step, I thought it was the case to ask if that alteration deserves a new question on its own. – burton0 May 29 '13 at 8:46
• Just an FYI, since you're somewhat new to meta--don't be worried about the downvotes on the question, as they don't mean the same thing as on the main site. On Main, they mean that the reader doesn't think the question is a good question. Here, they can mean that people disagree with the idea presented, not the way the question was asked. I'm saying this because it freaked me out when I asked my first question. :) – apnorton May 29 '13 at 12:31
• Thank you for clarifying this point. The net effect has been more or less the same on me :) I'm still not sure on what to do. In general, I agree that a new question has to be posed when new details are added. However, the impact on my alteration is so low that a new question may be redundant. – burton0 May 29 '13 at 12:39
• This is irrelevant to you, but just in case this discussion is found by some future user with a related question I want to add that while I agree with everything that others have said I would allow the question to be changed significantly, if the initial exchanges (say within a few minutes or so) with OP reveal that a key assumption was omitted. This is doubly ok, when the question became trivial due to that omission. After all, in that case the "wasted" effort was also minimal (and people worthy of their diplomas figured out that something's missing anyway). – Jyrki Lahtonen May 30 '13 at 20:02

This is not good practice.

While I can understand that you are reluctant to ask follow-up questions, think about it from the point of view of the answerer and the community.

The answerer has kindly taken the time to provide an answer to your question, and would be confronted with a changed question that renders the answer incompatible. Beside the offence the answerer could take from this, we have to consider the following:

• Either the answer will be adapted, in which case the answer to the simplified problem has disappeared for any future visitors to see, learn and appreciate;
• Or, it won't be, in which case the answer is useless to future visitors as well, since it does not answer the question any more.

In conclusion, after receiving an answer to the simplified problem, I suggest that you first try to solve the follow-up problem yourself (if you already tried it before asking the simplified question, perhaps the answer will help you to solve the follow-up problem), and otherwise, simply ask a new question -- linking it to the simplified version if this is clarifying or may help to express the desired amount of detail in answers to your new question.

• Can I do the converse, that is, forward referencing the new question in the old one so that the people who originally answered may be involved also in the follow-up question? – burton0 Jun 20 '13 at 15:04
• @burton0 Yes! I highly encourage creating a two-way reference. A short notice at the end of both questions which contains a link would be a good method. – Lord_Farin Jun 20 '13 at 16:12

It is highly recommended that one does not significantly alter the content of a question once a satisfactory answer has been provided. Doing so may make already posted answers irrelevant and perhaps nonsensical, and those that have provided these answers may very well feel annoyed at having wasted their initial efforts (and perhaps also be less likely to provide you with answers in the future).

Note that even if you do alter a question in this manner, community members can (and in my opinion should) revert your edits.

When you post a question it is understood by the community that that question is what you want an answer to, and the users will work to provide answers to that question. Subsequent questions should be posted as new questions.

It would be vastly preferred by many, including me, that your more elaborate version of a previously "easy" question be asked as a separate question; moving the goalposts during a game is not too nice to do. What you might consider doing, when asking the more elaborate version, is to link back to the "easy" question, and maybe explain why you have added the elaboration.