In a research interview, an interviewee once said something which caught me off guard, and to which I should have given a much better response. I was talking to a young and visibly Muslim woman (hair covered, but no more) about Islamophobia. Along with the usual instances described - name calling, threatening gestures, veil ripping and so on - this interview mentioned an act of omission, and one I disagree is Islamophobic. She said that when people she meets don't ask her about her religion, that is a form of Islamophobia. While I didn't respond at the time, what I should have said is that if I asked her about her religion that would be Islamophobic, or at least discriminatory, as in my day-to-day dealings with people I don't talk about religion.
This, of course, is not because I have no interest in religion. I do, but individual religiosities and practices are low in my list of salient topics. It's also something that we don't talk about in Britain anyway, like money and sex: it's not seen as polite. Further, I can look up information about how a religion works as an institution in a book or online, and it's likely that this source would be more authoritative than the individual. I could ask people about their individual experiences of religion, but as a life-long atheist they wouldn't make sense to me: I can't comprehend. Instead, if I'm meeting someone, becoming friends or acquaintances, I'm approaching the encounter with my salient topics... work, music, culture, comedy, global culture, football, and they theirs. In a dialogue, it seems natural that conversation will cover those topics that have a degree of salience for both. So with the interviewee I did end up chatting about work and culture, but not football, music (my topics) or religion (hers).
This makes me think that much of what society says about identity (and community) starts from the wrong premises. I agree with Amartya Sen in his Identity and Violence: we are tending to put all the focus on one or two aspects of identity - a combination of what we call religion, nationality, ethnicity, and 'race', which I end up calling ethno-religious - and make everything else secondary. In essence, there needs to be no necessary relationship between the official, national, or societal model of identity, local versions of same, and individual interpretations. An aspect of identity can be important in some circumstances and not in others. It can also be perceived as unimportant, while having effect: most obviously, if most people have X and this is not challenged by anything, then X is normal, unproblematic and invisible. There are also elements that do not compute - asexuality for example - within the standard understandings of how a particular identity works. Thus, trying to work out some kind of universal scale or framework is nonsense, as we don't necessarily agree on the initial premises.
This is not to say that we can't find patterns and structures. I'm sure that the status of Islam as a pariah religion makes my interviewee feel her religious identity more strongly. However, in a dialogue I have to be myself. I'm not immediately interested in religiosity or religious identity: I know people who are observant Christians, Jews, Muslims and other (although I also know a lot more unobservant/atheistic/agnostic people with backgrounds in all faiths!), but I don't ask them about their Sunday morning, Friday lunchtime or whatever. If I started by asking Muslims, then I'd be Islamophobic.
 I use Islamophobia as I'm following current practice, but I'm not sure we shouldn't be using anti-Muslim racism instead.