MaxPhysicalAddress is more readable than all the above. Pike simply had it wrong, IMO, probably from having been immersed in a culture of poor readability for so long that he had become blind to some of it.
I agree with Pike on this one - and I think it's likely true what he says about being used to reading prose. For those of us who do a lot of non-programming reading, Funny Out-Of-Place Capitals interrupt the smooth flow when reading variable names.
I do_tend_to_use underscores, though - I find that they don't have the semiotic noise value as WeirdCapitals.
I find the opposite — underscores are hard to read, but innercaps are almost effortless. At any rate, they're equivalent — you can mechanically transform from one to the other without data loss. I would say that either is preferable to not delimiting your words at all. It sounds to me like you agree with me more than Pike, modulo an insignificant matter of taste.
In 1989, external symbols were (possibly) only unique wrt the first 6 letters. Trying to guess if anybody else would name a variable MaxPhy* isn't a fun game...
That's true, but Pike didn't explain it that way. He wasn't talking about compatibility, but readability. That part of the article was explaining that longer and more complicated is not necessarily better in terms of readability, and there is a sweet spot where something is concise but not overabbreviated. That's all obviously true. Then he suggested that "maxphysaddr" is in this sweet spot for the idea of the highest point in physical memory that can be addressed, which I think is wrong.
