A panda walks into a café. He orders a sandwich, eats it, then draws a gun and proceeds to fire it at the other patrons.
"Why?" asks the confused, surviving waiter amidst the carnage, as the panda makes towards the exit. The panda produces a badly punctuated wildlife manual and tosses it over his shoulder.
"Well, I'm a panda," he says. "Look it up."
The waiter turns to the relevant entry in the manual and, sure enough, finds an explanation. "Panda. Large black-and-white bear-like mammal, native to China. Eats, shoots and leaves."
In my natural parsing, if you are enumerating things, once I hit the "and" then I know you have moved on to the next item. So, for me it is not needed - but that may be "British laxness" [1].
The author addresses this: "contrived sentences that mean different things depending on the placement of commas, because this tactic was so successful in fixing that thing that time."
The author conveniently dismisses sentences that contradict his premise as "contrived." That's not addressing anything, it's just employing the No True Scotsman trick.
Yes but contrived sentences are bad examples. If you have to carefully parse a sentence's punctuation to come to the correct meaning, the sentence should be rewritten. (Analogous to forcing code maintainers to remember obscure precedence rules to understand your code.)
I'm not sure what you're point is, becauae on first read it looks like you're agreeing with me? The whole point of the quote was that contrived sentences aren't good examples.
The point I took from that quote was that many punctuation rules come about to address difficult to parse situations. My point is that, if it's a difficult to parse situation, you should probably rewrite it in some way so that punctuation isn't so critical to understanding.
I will grant that the first form is clearly [0, 1, 2], while the second form could be construed as [0, [1, 2]].
HOWEVER, given that there is no "and" between 0 & 1, and that 1 & 2 are not like things, I believe most readers will naturally interpret the list as [0, 1, 2] and not [0, [1, 2]]. There are enough other cues that risk of misinterpretation is very low. Anyone who does interpret as [0, [1, 2]] is probably being willfully obtuse.
I think it largely depends on if the last two items in the list are a natural combination, either in the same category, or complementary things. The former is why you need the Oxford comma for "the strippers, JFK, and Stalin"; the latter is why you need an Oxford comma in "we brought the tent, bread, and butter".
Off-topic, but I like this version of the joke better:
There are two hard problems in Computer Science: cache invalidation, off-by-one errors, and cache invalidation.
But sometimes you need it. Here's an example:
"There are two hard problems in Computer Science: naming things, cache expiration, and off-by-one errors."
Take out the comma and you ruin the joke - at least to my ears:
"There are two hard problems in Computer Science: naming things, cache expiration and off-by-one errors."