RE: Does courtesy following really work? And is it Good Practice?
This COULD be a good idea - if it wasn't for the SPAM-BOTS that are active on this platform. You have, no doubt, seen some activity from bots - as there is already one bot that has posted in your comments so far: steemitboard. This is an example of a good/useful bot. However, not all bots are all that useful.
For example, take my post about "goodbot."
https://steemit.com/steemcleaners/@bi5h0p/everybody-please-join-me-in-flagging-goodbot-because-it-is-a-bad-bot
Now, to be fair, my opinion of @goodbot has been slightly revised since writing this article. I don't see it so much as a bad bot - I'm more neutral about it. It does have it's uses.
However, what you are suggesting will only make everyone who creates bots just go crazy with bots that comment on whichever posts they want to keep at the top of the cache - thereby totally defeating the purpose of what you are proposing, as well as spamming the hell out of everybody.
It's a very serious problem, and not easily solved - I wouldn't think - unless HF20 can also include some type of written-in code to prevent the bots comments from being considered, at all. Otherwise, we'll just have one GIGANTIC SPAM-FEST on our hands. o.0
So, at this point, I would like to introduce another idea into the SteemIt platform - which I'm sure has already been talked about somewhere: MANUAL CURATION ONLY. No upvoting-bots allowed. I hear there are "Guilds" of bot creators that regularly "conspire" to rape the daily rewards pool by upvoting each others' content with legions of voting bots. No doubt this has already been going on for some time, and TPTB are well aware of it. No doubt it has already been hotly debated before, in other posts, on other platforms, and behind closed doors where minnows do not swim. However, if SteemIt is to succeed in the long term, I see bots as a bad thing - unless the whales come up with some way to closely regulate their use and creation.
The thing is, I already know one thing about myself: if these bots piss me off enough - I have the ability to learn how to create and unleash my own bots. It will take a Herculean effort on my part to do so, and I'd probably rather be doing something else - but I have this excess reserve of manic energy that just dwarfs many others, by comparison. I do my best to direct that energy in positive and constructive ways, and I'm getting much better at it with the passing of time - but not everyone with these abilities is as magnanimous as myself. Herein lies the problem.
You are right, dishonest people will always try and find a loop hole. But what if a human verification was applied when commenting. It may make raise the quality of comments as lazy people would nit be bothers with their 2 word comments they then go and upvote themselves.
That would require Manual Curation Only. Sadly, we do not have this right now. I don't know if HF20 will address this issue or not - or even if they could. It would be nice, though!
https://steemit.com/steem/@l0k1/introducing-smackdown-kitty
I think this one might be something that fits the bill for what you are looking for to counter spam/circlejerk bots. Please do visit https://steemit.chat/channel/steem-coop-general if you want to talk with us about this particular issue.
Of course at present the development process has stalled, but as @personz, the guy doing the code for the bot, has been explaining, we are going to use it to attempt to neutralise some of the highest paying self-voted comments from users over 1000SP. To do more than this, it needs interested users to delegate SP to boost her ability to smack these wankers down (and I don't apologise for making an analogy between self voting yourself a $1+ reward on a comment).
Why I am referring you to this is, perhaps you can make some suggestions how we could expand the scope or refine it somehow so that we can also have her targeting these circle jerk bots, or lay out more clearly how you see the problem and what you think can be done about it. I am very good at exploring problem domains and discovering attack surfaces, so perhaps we can help each other.