By having a shared library they will have the weapons to tap into a piece of the market who doesn't see Nintendo's home consoles as a good alternative to their tastes.and would never buy a handheld for games they consider best on a big screen.
Heck, it's no alternative since Xbox came into the gaming market.
Sony fans will choose XB1 as an alternative before they will consider buying a Wii U as second console.
Shared libraries will allow for development costs (even upscalling costs) to be diluted on what they can sell on the other console.
For example, Pokémon will always sell more on handhelds, but they can also get sales on the home console front from people who, despite being interested in the series a) would never buy a handheld for that, b) would never buy a handheld, period.
Shared libraries will allow, with today's expensive game development, for games like W101, Bayonetta, etc., to make a comeback or even being developed just because the userbase would be larget by default.
By sharing libraries it will mean Nintendo has, in that option, the most resources to develop games that XB and PS fans actually crave. And if they succeed, 3rd parties will see that there's a market for them.
Sharing an OS is a short term solution to speed up game development. You might spare time by learning one code instead of two, but you won't gain that much time to allow for a significant number of titles to be developed with one code instead of two.
Nintendo might lose consumers that want both a handheld library and a home console library, but they will gain in software profits (one game for two consoles instead of just one) and those who would buy a home console if a certain type of library existed on home consoles.
How will Nintendo convince 3rd parties that its userbase will buy their games if the userbase is filled just with Nintendo type of games?
It's not 3rd parties that will be able to create that market and then cater to it.
