I'm a former student of his. He was one of the people that made me from a teenager that hacked on his keyboard to get something to run to a seasoned programmer that thinks before he codes.
Even before I met him at the university I was programming in Oberon because there was a big crowd of programmers doing Wirth languages on the Amiga.
I'm also a student of his, and later met him socially on a few occasions as a graduate student (in a different institute).
Undergraduate students were all in awe of him, but I got the impression that he did not particularly enjoy teaching them (Unlike other professors, however, he did not try to delegate that part of his responsibilities to his assistants). He seemed to have a good relationship with his graduate students.
In his class on compiler construction, he seemed more engaged (the students were already a bit more experienced, and he was iterating the Oberon design at the time). I remember an exchange we had at the oral exam — he asked me to solve the "dangling ELSE" problem in Pascal. I proposed resolving the ambiguity through a refinement of the language grammar. He admitted that this would probably work, but thought it excessively complex and wondered where I got that idea, since he definitely had not taught it, so I confessed that I had seen the idea in the "Dragon Book" (sort of the competition to his own textbook). Ultimately, I realized that he just wanted me to change the language to require an explicit END, as he had done in Modula-2 and Oberon.
Socially, he was fun to talk to, had a great store of computer lore, of course. He was also much more tolerant of "heresies" in private than in public, where he came across as somewhat dogmatic. Once, the conversation turned to Perl, which I did not expect him to have anything good to say about. To my surprise, he thought that there was a valid niche for pattern matching / text processing languages (mentioning SNOBOL as an earlier language in this niche).
No, Borland did have a Modula-2 compiler (where actually Martin Odersky of Scala fame worked on), but they decided to focus on Turbo Pascal and sold it.
I'm not suggesting Turbo Pascal was written in Modula2, I'm saying it implemented Modula2, not Pascal. Modula2 is a superset of Pascal. Pascal never had modules AFAIK but Turbo Pascal did.
At least several Pascal, Modula-2, and Oberon-2 compilers.
My very first compiled programming language was Pascal. I got the free "PCQ Pascal" from the Fish disks as I wasn't able to get the C headers from Commodore which I would have needed for doing proper Amiga programming. Likewise later Oberon-A although I don't remember where I got that from.
There were also commercial Modula-2 and Oberon-2 compilers. I just found that the Modula-2 compiler was open sourced some years back. https://m2amiga.claudio.ch/
Even before I met him at the university I was programming in Oberon because there was a big crowd of programmers doing Wirth languages on the Amiga.
He will be missed.