The 4th poll submission isn't working for me for some reason.
But then I really don't want to answer that one anyways. I liked the article, but I really disagreed with the polls at the end. I kept thinking over and over, "This isn't how it works at all." At least in my games. Maybe for others, but the answers simply don't apply so easily to how we play.
So here are my answers in written form for FWIW:
1. Simplicity - Both for players. As high as he or she wishes to manage for the DM. In general the higher the better for the DM, but great elegance works best. For players it's all about how much they wish to invest. We have about a dozen stats plus purchasing. Half of the previous are rolled for and the others are a pool of choices (like Race, Class, Alignment) with custom normally a last option. But for massive customization extensive background creation is possible, even throughout the campaign between sessions. All of that is turned into mechanics by the DM. The player's do not need to know how, they simply are tested on remembering what they did write. So both simple and potentially extraordinarily complex for player in my book.
2. Quick creation is essential for when a player loses a PC and wants to start a new one mid session. This feeds into #1 in that further customization can be done afterwards, but quick-starting into the game is easily and pretty much alone. Continuing higher customization and complexity can be written or generated for later sessions - the current detail only needs to cover a few hours anyways. What changes is when players want to use the first session as a party creation process, which can really amplify the game. By working out who wants what and how each can support one another the players are already working together as a team. This means detailed character creation with all the interconnections between PCs set up by the players to the degree they desire at start. So, for me, quick creation is necessary for mid-session play, but shouldn't be the only option when beginning a long campaign.
3. "Ability to Customize Fully" - This one's is sort of like the nature vs. nurture argument. It's about 50-60% nature or rolled generated abilities and 100% nurture / environment chosen customization of abilities. A PC is generated by chance rolls representing results players don't get to make. But then those results can be manipulated in the "background" by buying up and down different numbers. So that Dex can be raised (focused on in training) to qualify for the Illusionist class. But the rolls don't normally allow everything. However, beyond manipulating the information the player may also drop the character (which then becomes an NPC) and roll up another. The tough part for players comes from the choices. What class, race, equipment? Each choice improves the character in some ways and not others. But customization can also occur in each of these areas too. But "I want a Wizard who fights as well as a Monk hand-to-hand" doesn't start that way. In order for the choices to be meaningful there must be a tradeoff rather than flat out omnipotence. What I mean is a level 1 Wiz is less powerful than a level 1 Wiz + level 1 Monk in abilities. But customization at start could allow for a little of both. Customization through play could allow for both later on. Full customization is there, but influence is always limited - though game play is the promise that this can grow. [Unlimited influence, in-game omnipotence, is always the rejection of customization in the end anyways.]
4. "When it comes to character creation, what's more important to you, simplicity or customization?" This question doesn't really make sense given my other answers. Both really, it depends upon the point in the game and the level of opting in by the players.