With respect to claims geared to an end result, often you wind up with the same problem: criteria to determine when the end result is attained (function is performed). In your simplified example, the question then becomes when does the pipe reach from the first shelf to the second shelf? What if it's shy by 1/16 inch or over by 1/16 inch? Does it infringe?
Well, perhaps my example was too simplistic. The "reaching" should achieve some goal. Suppose the first and second shelves should support the pipe thereacross. Or suppose the pipe should "reach" such that a cue ball can roll from the first shelf, through the pipe, and on to the second shelf.
In theory (and hopefully in practice in the context of a real application), failure to achieve that goal in an accused device will deny significant value of the the accused device such that it's okay if it doesn't infringe, even if it fails to allow the cue ball to roll to the second shelf by only a millimeter or even less.
Similarly, if I have a table with a top, a base, and four legs, I can specify the lengths of the legs such that the top is [substantially] parallel to the base. But the problem then shifts from measurements of the lengths of the legs to measurements of degree of parallelism between the top and the base.
Again, it might be worth trying to find functional measurements of level-ness. How about "sufficiently level such that a new, clean sheet of 20-lb bond paper of US letter size will not slide off the table top when consisting essentially of clean, unmarred formica?"
It might be the arts I work in, but usually things do what they're supposed to do or they don't. How "level" a thing must be to be a table rarely, if ever, comes up for me. However, some things I'm working on now will probably have those "how much is enough" issues. Ugh....