I'd tend to argue that it's more important to keep clearly rejected suggestions than it is to keep ones that have been both accepted and acted on.

@Glen_b I can see your point. My reply is that I would like the synonyms thread normally to have 0 (non-deleted) answers; once a new answer appears it gets discussed and eventually acted upon or rejected. We have recently had over 10 (maybe it was 15 or so?) active answers there and I think this makes the thread crowded and difficult to navigate.

If we undelete all the rejected suggestions (around 10), it will make the thread very crowded again, and not very usefully so. I think this will directly intervene with people's willingness to look through the thread and comment/vote.

That is why I think deleting is a good idea. However, I can see your point about keeping a record of what was rejected (in fact, keeping a record of what was implemented can also be useful, even if less so). Maybe we can come up with some way of keeping this record? E.g. we could make a wiki answer in that thread to keep the list of suggestions.

I don't like this particular solution too much because this wiki answer will become long and will be obstructing pending discussions. Maybe there is a better solution (adding a list to the question itself? starting a separate Meta thread?)

@amoeba Yeah, I understand. This is one of the problems of aggregating all these into one thread (the advantages are obvious enough). Possibly a better alternative might have been to have one thread as the organizing thread but linked through to individual questions in their own posts; deletion in the organizing thread would leave it only holding current information but the old votes could still be found. That would probably involve too much work to maintain though.

Okay, you're right that deletion is better than the alternative, so I think there's nothing to do now.

@0x90 It's a nice question, but it doesn't seem to have anything to do with statistics or machine learning. As far as computing the derivatives of a vector field goes, math@se would be the site to ask that question. And although you are a Matlab user, I'm sure if you were to pose your question on Mathematica@se you would get excellent and varied suggestions about how to visualize 3D vector fields.