Re: The problem is implementation
First off, they bought Spinnaker in 2003 and 2004 or 2005 came out with ONTAP GX. IIRC around 2008/2009 they were the first to shatter the 1 Million IOPS Mark in the SpecNFS benchmark with a 24 node ONTAP GX system (which, according to reports at that time, was not a lab queen, but deployed just like that at multiple customers). I really don't know, where you get your 14 years from...
Second, it really depends on your definition of scale-out. It's not a shared nothing system like Solidfire, true. But if you can nondisruptively grow capacity and performance, e.g. using FlexGroups up to 20+ PetaBytes and 400+ billion files in a single namespace, by adding controller pairs, that's scale-out to me. In the end it's not about semantics, but about customer outcomes. And again, NetApp has a history of doing things differently (I remember "that's not a real SAN, they have a file layout underneath" from ~2009).
(Not a NetApp employee, I just teach storage for a long time...)