The reason I jumped you is because, in your usual style, you attacked the approach taken by someone else just because you didn't like it and didn't understand it!.
I think the lack of understanding is on your part. In my most recent post, which I think you did not read before posting yours, I realized that the probable "reason you jumped" was that you did not realize that in programming the chip in question, feedback (transmission from the '51) is involved.
You have, in this discussion which ended up being about input again and again referred to the requirement of a bi-polar output which the post where I "in your usual style, you attacked the approach taken by someone else just because you didn't like it and didn't understand it" reacted to to urban legend that you can safely transmit RS232 with "a npn transistor". As far as "my usual style" I always react when someone post a marginal solution to a problem as if it is 'perfect'.

It would be both interesting and informative to see a supporting reference, ...ANY... supporting reference, to the "science" to which you refer.

I'll skip the "supporting reference", the below should be enough.

If you have a 6V noise margin a noise pulse of more than 6V is required to tilt the load.

If you have a 2V noise margin it only takes a 2v noise spike to tilt the load.

Erik