For some perspective, this was an app developed by 5 core devs that grew to ~125M+ users who by necessity were split across implementations concurrently released for Windows 3/95, NT, OS/2, Mac, Motif, OpenLook, because we were an early enterprise product and needed GTM help offered by each of those platform vendors.
Our tiny dev team chose to prioritize platform breadth thru a "portability layer" developed in a pre-internet pre-browser era, with the constraint that its ~3M+ LoC (in C) had to run in a 250KB working set that grew to a couple MB by the time I left.
Yes, I wish the UI could have been better on each platform. But as a pragmatist I remain proud of the paradigm and its usability given the constraints.
(see also Ray Ozzie answer in other comment about portability layer https://news.ycombinator.com/item?id=26697054 )
Notes issues were a nightmare, exceeded only by the problems caused by oversized apple talk and netbios networks, and physical layer problems in networks that predate the "base T" paradigm.
Things you got with Notes out of the box with minimal systems configuration effort: easy document sharing, easy authentication, entitlement, encryption of documents (and emails) across an organization. Easy provisioning of strong encryption across an organization. I could send digitally signed documents both within IBM and to peers outside IBM fairly easily. I could create a library of security processes for IBM's internet operations, pick out the organizations and people who could have access across the company, and it just worked. For bonus points we got document retention, audit logs, it all just was there out of the box. I didn't need to know ahead of time if someone was using Windows or OS/2 or a Mac.
It's easy to dismiss it all today as horrible bloatware, and as "just" an email tool it was probably overkill. I think too many companies (IBM especially) bought it, set it up, and utterly failed at educating people how to use it (IBMers were notorious for sending around email chains with multiple copies of Freelance or Powerpoint presentations, instead of a link to the presentation in a central repository, partly because IBM's services arm and IBM's CIO didn't want the overhead of managing such things.).