I do feel that in many circumstances there exists a gap in our definition of what is a "useful" comment that should be preserved. Chatting back and forth I think we can all agree belongs in a chat room because most of it is non-relevant or it involves two or more people discussing aspects of the question or answer that could hopefully lead to an improved question or answer, or possibly other useful questions. These types of comments tend to outlive their usefulness though once they lead to positive edits and changes in the question or answers and generally they can be removed.
I think though a common theme that we see frequently in the Software Engineering site is that students, junior level engineers and developers or generally less experienced professionals will run into a problem where they need some guidance, and they may have some specific question that they THINK they should ask, when in reality they are cluing more seasoned engineers into a more general problem that the OP should consider, or they have a fundamental misunderstanding of some important concept.
So when this happens we are really at a fork in terms of how the community will respond. The question itself may in fact be good and valid, but a more experienced engineer will see that even if they directly answer the question, they are not really going to help the OP with their most likely problem. The question and answers are relevant, however some very important and valuable comments can be made that are hugely beneficial to both the OP and potentially others in the same plight. Those comments again, may have nothing to do with the question or answers, but are potentially important.
I know personally that if I were to blindly follow the rules then I would be expected to delete such comments, but then I think that is why community moderators have some discretion here. It is inherently subjective, so you can't always apply objective literal rules to this.
The rules INFORM our decisions but do not dictate them.
I am wrestling with the idea in my head at the moment to propose a change on Stackexchange Meta where we can classify comments and potentially segregate them in a more meaningful way so as to preserve interesting and informative commentary without it taking focus from the question and answers themselves. I just haven't organized my thoughts on it and would like community involvement here to help me figure out how this can potentially work on a grand scale before I throw it to the wolves on Meta.
I do not enjoy contributing material that may be removed without warning, without recourse, and without reasonable, common-sense rules for determining when I can and cannot say something.
The thing that I want people to keep in mind however is that any content in any form that we provide on this site is community owned so it is at the discretion of the community if a question, answer or comment should be edited or deleted. In no way is anybody arguing that you SHOULD NOT say something that you feel is important. If it is not rude or abusive or violates site rules, then you SHOULD contribute if you feel you have something valuable to say. You are not doing anything wrong here but I understand that it sucks when what you post is not being received well by the community. Understand that someone in the community flagged your comment and a moderator reviewed and made a decision about it. Nobody is actively hunting for these things ( i would hope not ) trying to "cleanse" Stackexchange and certainly nobody is targeting your content purposefully.
Falling on your sword and playing the martyr isn't going to be helpful either as it just effectively encourages division and discourages communication on behalf of the community. Lets focus our passions to how we can solve this problem.