Before the first Table of Contents, who would have thought such a thing was necessary? Once someone real-ized the need (see Heidegger on techne in his "Essays on Technology") ToC came into being as standard.
Indexes? Maybe before, maybe after, but same process. Footnotes? Likewise.
What's fancy about a dictionary? But before Sam Johnson's effort (1755) there was no such thing. There were collections of words, sure, dating to the mid-1400s ... to help with spelling, but no more. And even those very few (in 300 years!) consisted of something like 5K words.
Before the first, the requirement was anything but obvious.
How about spreadsheet? Imagine having tried to explain what a spreadsheet does before spreadsheets existed. Without a working prototype to demonstrate, you'd be thought a lunatic.
Documentation is the corner-stone of literacy. Having it is no guarantee, but still: every "obvious" initiative was far from obvious before it was realized, so the absence was taken as "normal".
Case in point: 1998 I was discussing comment threading and nesting with Dan LaLiberte, regarding his HyperNews project. Some 5 years later when I talked about that I could point to an implementation, with LiveJournal (a recent example of nested threading).
So now, 2010, is that functionality standard? common? ubiquitous? How about this: 12 years after my discussion I find that this practice is still rare and exceptional. Go figure.
Let's just get it done!
p.s. I started working on my "participatory deliberation" system in 1975. To me, the need was obvious. That's what comes from having done SigInt for a living. *grin*