Mike Masnick at TechDirt would likely answer Never!
Martin also asks And most importantly, can the plagiarism of intellectual property ever be proved?
IPBiz, of course, notes that plagiarism and copyright infringement are two different matters.
Can infringement of copyright be proved? You bet. Happens all the time. George Harrison got fingered once. Ian Fleming had an interesting problem with Thunderball.
Can plagiarism be proved? If plagiarism is simply copying without attribution, then proving plagiarism amounts to showing copying has taken place and showing there is no attribution. In the Poshard Ph.D. thesis plagiarism, the students at the SIU Daily Egyptian laid out original and copied texts side by side. This is not rocket science. Poshard never said he didn't copy; he lapsed into a discussion of inadvertent plagiarism. Then, in following up, SIU copied, without attribution, its definition of plagiarism from Indiana University. The case is closed on SIU. Plagiarism proved, but it doesn't really matter, after all.
Martin talks about one of the more recent examples of plagiarism related to Harvard:
Kaavya Viswanathan’s How Opal Mehta Got Kissed, Got Wild and Got a Life – greeted with rapturous advances from its publishers in 2006 and swiftly signed for a film deal – turned out to contain passages with strong phrasal similarities to five separate writers, including chick-litters and Salman Rushdie.
Strong phrasal similarities? Readers of chick-lit spotted evidence of copying of text, showing that the readers were more familiar with the product than the editors.
Martin doesn't answer (or even ask)the obvious question: why didn't all those publishers spot the plagiarism on sight? Indeed, why didn't Poshard's thesis committee recognize that plagiarism on sight?
Martin writes of Kaavya: she retreated into obscurity. There are many who would love to retreat into the "obscurity" of Harvard.