Just read Alan Kay and Adele Goldberg’s Personal Dynamic Media: http://www.newmediareader.com/book_samples/nmr-26-kay.pdf This line really stood out to me:

(If) the projected audience is to be “everyone,” is it possible to make the Dynabook generally useful, or will it collapse under the weight of trying to be too many different tools for too many people? The total range of possible users is so great that any attempt to specifically anticipate their needs in the design of the Dynabook would end in a disastrous feature-laden hodgepodge which would not be really suitable for anyone.

I feel like what we got (modern desktop environments) neither collapsed under their own weight, nor became feature-laden hodgepodges, but we get most of our features and functionality from an arguable hodgepodge of independent isolated apps.

Regarding the bit about a hodgepodge of independent isolated apps, a lot of what divides apps and their data is just the way capitalism works being reflected in the platform. We have various roles, e.g., code editor, email client, and lots of different apps compete in that space against each other (Gmail, Superhuman, Outlook, Mail.app). Apps don't talk to each other, or use the same data structures, because they compete with each other. You can actually see theses walls breakdown when one company owns a bunch of apps, Microsoft Office, Adobe Creative Suite, and Affinity's apps all illustrate this with their interoperability features. Any view of a more seamless/monolithic desktop that doesn't solve this problem will never work, because if apps can't compete with each other than the platform will stagnate.

I like to look at this from the bundling/unbundling strategy perspective as I think it’s a more complete view of what is happening. Benedict Evans very recently published this which seems somewhat related: https://www.ben-evans.com/benedictevans/2019/9/27/new-productivity