As you say there is a huge world of difference between chopping up words to make up the solution and chopping up the clue to make it solvable.

First of all I ought to say that I don’t really mind the in-deeds but concede that they are the thin edge of what could be a very fat wedge. Where people draw the line along that is according to taste and I perfectly understand the concerns of people who don’t like it at all (and other than some people expressing a dislike I don’t think anyone here has been quite as hardline about it as Geoff makes out). Even if you think it’s OK you have to agree that there must surely be a limit as to how much pre-processing of a clue we should be expected to do.

For example would “Mingle as a star does” be OK for M in GLE AS for GLEAMS? Would “Blundered i.e. acquiring consumable” be OK for (BL under ED I)+E for EDIBLE? How far do we allow things to go. This is like having to put the clue through a blender before we solve it!

Finally, Paul B, I’ve definitely seen in-deeds outside the Guardian and am certain I’ve even seen them in the Indy (I think that the Indy is perhaps more liberal than you seem to think).

Whether or not such constructions are fair seems to me to be a slightly different discussion, one that does not necessarily equate to whether or not one holds Ximenean views. We might consider Araucaria for example (with his devious tactical array, not short of in-deeds), deemed unfair by the Dreaming Spiral Inquisition, who has by some peculiar quirk managed to become the most famous, most loved compiler on the planet. More loved even than Don, I am told (although personally I find this incredibly difficult to believe).

Consider too the compiler under scrutiny here, whose puzzles may be found across a range of papers. I am certain he deploys these supposedly unfair devices in The Guardian simply because they are allowed, and not because he thinks them justifiable. In his work elsewhere, I do not think you will see them.

It’s not about an unacceptable device somehow becoming acceptable as, say, a grammatical cosntruction might. It’s about the clue saying what it means. To me ‘X in fest’ says what it means when the intended result is ‘fXest’ while ‘X infest’ doesn’t say this at all.

This is an unsuitable forum for lengthy discussions on this point but I really would recommend Azed’s, Don Manley’s and Ximenes’ books for deeper discussions of this idea.