I think my attitude to the reader-writer contract is pretty standard. A reader looks at a text and decides on an initial reading strategy, taking into account genre, length, shape, reputation of author etc. Experienced readers are more likely to adjust their strategy, or adopt a sub-strategy. Readers may give up if their strategy isn't working. They may decide that the piece isn't good [of its type - the type the reader perhaps wrongly assumes] or that the author has deliberately misled them. They may find the deception irritating or playful.
Sometimes authors give advice on how to their texts might be read. In her recent "More than Weeds" L.Kiew writes "I do not italicise words because none of them are 'other', 'exotic' or 'foreign' to me ... I believe that all comunication is to some extent partial and problematic; and poetry is to me one of the least dogmatic of the artforms ... readers are curious and able to enjoy the sounds and shapes, to dig out meaning from context, and to explore using the many tools and resources [that] are available online". I like it when authors try to help in this way - I don't think it "collapses polysemy", it makes me trust the author and read on. But I don't think writers should assume that readers will be curious rather than irritated.
Of course, a book might succeed despite an author's intentions (a piece meant to be taken serious may best be read as a comedy) or the advice may be part of the game.
I use obscurity as a device, and I've some understanding about the uses of obscurity (see my poetry and obscurity article for example) but over the years I've come to distrust authors more often. I'm less willing to battle through obscurity if I see no purpose in it other than trying to mask the author's inadequacies - i.e. if I think the author without aesthetic loss could have reduced the muddle, I wonder what their game is. I'd like more authors to appreciate the disadvantages of using obscurity - e.g. that readers might stop reading, might think the author thoughtless, elitist, or rude.
Amongst the newer examples of obscurity I see nowadays is when in the same book a poet uses various alternatives to line-breaks, and sometimes uses inline spaces instead of commas. If a poet makes readers think that there's a purpose (meaning) to something, the poet shouldn't be surprised when readers are frustrated to discover that there is no reason why "/" is used in one poem, "|" in another and line-breaks in another. Poetry layouts can all too easily become obscure - even good old line-breaks are often puzzling enough.
For some other viewpoints see -