Questions about suggesting reviewers

In our July "how can we help you?" thread, two readers asked questions about journals asking authors to suggest reviewers:

I work in philosophy of science, but recently I have started to submit papers in another subfield of philosophy. These are all journals with a good reputation. However, to my surprise, it seems that a pattern is being established: editors ask me to provide suggestions for possible reviewers. This never happened to me in philosophy of science. What's the best practice here? Naturally, I'd propose people I know (assuming they have not read the paper and they do not know I'm working on that specific topic), because I know that those people will take their job seriously. However, even if those people do not know that I'm writing that specific paper, they may recognize who I am. But this, I say, happens even in blind peer-reviewed, especially in highly specialized fields such as philosophy of science. Am I terribly wrong? - Posted by: Don't really know what to do

I have a question related to that of Don't really know what to do. Should a philosophy journal ask the author of a submission to suggest possible reviewers? What are the rationales for and again this? -Posted by: jack 

Another reader answered:

This is a common practice in science journals, to ask for (i) possible referees, and (ii) sometimes a list of people who you do not want to review your paper. In science it is common place to have only single blind - so the reviewer sees who the author or authors are. I publish regularly in an empirical field (call it scientific), and the key journal will not send your paper out for review unless you have your name on it.

However, while this may be a common practice in science, I wonder how readers feel about it, and whether anyone has tips for 'best practices.'

For my part, I've had to do this as an author several times recently, and I have to confess that I feel a bit uneasy about it. Although I did my best to suggest reviewers with integrity (viz. specialists in the relevant fields rather than friends of mine), there nevertheless seems to me to be a clear conflict of interest here: namely, the fact that authors have obvious incentives to suggest reviewers they have some reason to think may be sympathetic to the paper, or even reviewers they suspect (or know) will recognize them as the paper's author. Finally, although I'm not entirely sure, I may be a bit less uneasy with a similar practice that I've also recently seen: journals asking authors to propose reviewers to avoid. Yes, a similar conflict of interest surely arises here as well. At the same time, if an author has grounds for thinking that a particular individual cannot review a paper fairly (perhaps because the person has seen the paper before, perhaps at a colloquium or conference), then it may make sense to alert the journal to that fact, so as to preserve anonymized review.

What do you all think? Should journals engage in these practices? And in any case, how do you think authors should go about suggesting reviewers and/or reviewers to avoid?