I'm speaking to this discussion, not the submitted article. Sure, start with a web app (though it really depends on the target -- if the web app yields a sub-standard interface, performance, or utility it's a dead end) with hopes of building an app when you have an established base.
However in this discussion there are a lot of people who are arguing from the perspective of a world that doesn't exist, based upon the same "this changes everything" argument we've heard every year.
However in this discussion there are a lot of people who are arguing from the perspective of a world that doesn't exist, based upon the same "this changes everything" argument we've heard every year.