OOP as originally articulated by Alan Kay is not the same thing as what's represented by C++, Java, et al.
The original idea was that since a computer can simulate anything (including a better computer), that computers should be the smallest things you would want to recapitulate in a system. These computers (objects) would communicate by passing messages. Each object is the interpreter for the messages that are sent to it (similar to how your laptop interprets messages you send to it via the keyboard, or how another computer communicates to it through a network interface card - objects are also parsers). Crucially, in real OOP, objects are Turing-equivalent computers, and systems are composed of objects that communicate by passing messages. You're using an exemplar of such a system right now.
In fact, Dr. Kay has lamented that he didn't call it something like "Message Oriented Programming" since messages were the actual big deal.
I will grant that I should've spent some time coding in Smalltalk by now to get a better sense of the implications of that style. I myself find it surprising when people delve into a language debate pitting a traditional imperative/OO language against an ML without having knowledge of it.
I still don't think it's fair to celebrate Kay for OO then being that it's not what he intended us to end up with, so certainly he must be recognized for something much more relevant?
Other than the huge list you can get from Wikipedia, the one that blew my mind recently was STEPS, a research project he was involved in a few years ago:
Here's a long vid of Kay showing off the fruits of STEPS and doing his job to remind us that because we forgot about the work done in the 50s-70s, we've spent the last 40 years rediscovering it and are only recently starting to catch back up.
It's important to note that he is giving his presentation using STEPS' custom productivity suite; running on STEPS' custom OS; both of which were compiled using STEPS' custom languages and custom compilers all the way down to the assemblers. And, the whole package (OS, compilers and all) totals well under 100K lines of code with no external, non-STEPs code introduced.
I had seen that video before... I found it upsetting as most of it was bragging and ranting, am I the only one who noticed?
There was no material in that video that anyone could use and following the above article (http://gagne.homedns.org/~tgagne/contrib/EarlyHistoryST.html) also was written by Kay himself (which I haven't read yet but will do)... I'm starting to wonder whether the reputation is self made.
STEPS looks like a new project from 2006 with this 5-year progress report published in 2011... I can hardly believe Kay is recognized for that being that it's far too recent and none of it is mainstream.
But here, I conceded, I don't know. Please enlighten me.