In it, its authors (Nicholas Economides of NYU’s Stern School of Business and Brian Viard of the Stanford Graduate School of Business) claim that it pays Microsoft, or any creator of both a required base product, and a complementary product that runs on that base product, to price the base product relatively low, attracting others to also build products that use the base product and create a large ecosystem.
That, of course, is exactly what Microsoft does. In the article, its authors estimate that at the time of the antitrust trial Microsoft sold Windows for a street price of about $60, while sellling Office for about $240 (or about four times as much).
The existence of many additional players with other complementary products on their base Windows product increases the value of both the base product and the complementary product, they claim. They can make up any losses created by the less than value pricing of the base product by charging more for the complementary product.
Since improvements in either product are valuable to the vendor in both products (and presumably show that value in increased sales), the vendor (in this case Microsoft) is more likely to make such enhancements over time. This may sound counterintuitive to you (it did to me, at first), since we are inclined to believe that monopolists are disinclined to enhance their products.