So in discussions about LJ, people have their opinions about assorted features and the people who use them.
There's this great little pattern where someone states their absolute dislike of a feature, generally based on too many people doing it wrong, and people who use and like the feature will get drawn into a (often heated) discussion about how good the feature in fact is, and how they personally are Doing It Right.
The obvious part of the problem is when the complaint is in fact addressing the user, and not the behavior or (best of all) the feature. Consider the following comparable statements:
I hate people who set their background to bright pink.
I hate it when people set their background to bright pink.
I hate backgrounds set to bright pink.
The first is a clear invitation to an irritated slapfest, as people who have their backgrounds set to bright pink may wind up insulted, as whether the complainant has intended to or not, they've just said, in effect, that they hate those people. (Bonus points when the people with the bright pink backgrounds are in fact friends of the complainant, as this may be interpreted as the declaration of the end of a friendship.)
The second addresses a behavior, but it's still a thing that people do that's being complained about, and not a technical feature. This naturally devolves into "Don't read them, then!" rather than a technical argument.
Only with the third does everyone have the best chance of focusing on the technical problem at hand, that of making sure that this particular user does not see backgrounds that are set to bright pink.
A less obvious cause is when the people who use and like the feature fear that someone else's dissatisfaction with the feature will result in that feature being taken away. People who love the feature want to make themselves known, sometimes to the point of wanting to drown out the voice of the person who does not like the feature, to make sure that no one in a position of power makes the mistake of thinking that no one likes and uses the feature.
This is simple to address, given a spokesperson in a position of power and accurate server-side statistics. Someone in a position of power merely has to pull out some statistics on actual use of the feature, and state that there is no danger of the feature being taken away, but they are all for making the feature generally better: better for those who like it, and less obnoxious if possible for those who do not. People who have not got to that part of the discussion yet will still raise an alarm, but it can be pointed to in order to direct the debate more helpfully.
People who use and like the feature may have taken the original statement about the first party's dislike of the feature as a value judgment about them, personally, or a class of people like them. This happens no matter what way the statement has been worded, but is always worst when the phrasing is about the people using the feature rather than the feature itself. In any case, people who use the feature are likely to be sensitive.
People, in general, do not like it when others dislike them. It's unpleasant to find out that something that you like to do is irritating others. If someone who uses the feature sees that someone is complaining about people who use the feature, it's hard to let that slide without some sort of reaction. (A particularly volatile situation can arise when two friends, one of whom loves a particular feature, and the other of whom loathes it, discover their disagreement of opinion in the same thread.) Sometimes the reaction is helpful. Sometimes it is not.
The culture war between those who love and hate a particular feature comes up a lot in feature discussions, and can take a discussion about a feature into a related but not entirely topical discussion about how to use a controversial feature "correctly", that is, in line with the unwritten social contract between users of the site. Some of these discussions are very useful, but a technical forum is hardly the best forum for the discussion; some of them devolve into slapfests and mutual dissatisfaction.
A common reactive response to "I don't like it when people I read do X" is "Well, don't read people who do X, then!" People who say this a lot should be slapped around a bit with a wet fish. Attention is the primary currency of an online community, and, indeed, social interactions as a whole. Proposing that someone completely shun someone they are interested in interacting with as a response to a petty annoyance is complete overkill. Yes, one of the possible responses to irritating online behavior is avoidance of the people exhibiting the behavior, but it's hardly the only social response, and should not need to be the only technical response.
Unfortunately, LiveJournal's features create an environment that reinforces a cultural tropism toward that extreme. The friends page structure, which has an all-or-nothing approach to entries by someone (on or off the friends list, in or out of default view, in or out of another custom friends group used to filter the friends page) leads to a culture in which partial ignoring of selected entries while maintaining a reading relationship is deprecated. Partial avoidance on the part of the reader is seen as a cop-out: reading must be all or nothing.
If the writer has chosen to do something that irks or inconveniences the reader, the reader may or may not actually have have tools at their disposal to deal with the situation. There are image placeholders for potentially not-safe-for-work or page-breaking images, for example. There is ?style=mine for journals that display unreadably to the viewer. On the whole, however, there are more tools for the writer than for the reader, such as the lj-cut, or custom friends groups. Thus, most of the social burden of choosing an appropriate audience falls on the writer.
A discussion about how much someone is irritated by others' use of a feature and would never like to see it should be guided into two particular productive avenues: first, what specifically about that feature irks them, and how it could be made to not irk them when they are presented with it; second, what technical tools could be made available to them so that they are not faced with it. Technical tools for the writer to avoid irking the reader is another possible productive spin-off.