The conversation from @gtg has a lot of the reasons for how likes are complicated.
I believe what you are searching for is as he describes a "front end solution" and would likely not be decentralized among the many front ends.
As stated simply counting any vote as a like would result in manipulation of the data by someone using hundreds of useless accounts.
Also on top of it any feed of posts requires an API to give that sorting order based on the proposed "like algorithm".
Someone for example could create an algorithm and either convince all nodes to provide that algorithm OR run their own steady functioning API to produce the results of that algorithm whenever any user requests to see a feed based on that algorithm.
But then we go back to needing a system that can track LIKES in a way that isn't so easily broken it doesn't brake the whole system.
This basically rules out use of VOTES alone because votes are designed to fully allow automation. A front end perhaps could check when the action is done manually and not automated... that's my assumption maybe developers can verify that assumption (cc @asgarth) Or the front end could have a separate button clicking system with a symbol like a thumbs up. But then the front end would have all that data and would have to do it's own work checking for manipulation and probably keep how and why manipulation is working against a person under-wraps in order so that the manipulater doesn't know how to bypass their manipulation attempts.
A system to count likes from a front end is probably pretty straight forward the system to do self-checks and look for manipulation is the most complex part of it and in the end a system like this is basically not fool proof and accepts some manipulation as part of the package... which isn't the end of the world because when people aren't earning money and maybe only sorting preference (for just those who choose to sort content by this algorithm) then the abuse of this is likely not very high to begin with and probably for quite some time until placement of posts becomes worth the effort to beat the self-checking of the system.
Adding a like button to peakd is the easy part. Probably even tracking which votes are manually given on peakd itself. However presenting a feed for a topic page or the ALL TOPICS page is where it starts to get more complicated.
THE BEST SCENARIO
On another note this whole concept probably works easier in a community setting.
Counting the VOTES of each added MEMBER of a community... that way the community moderators are doing the job at qualifying who's votes should actually impact the algorithm.
thanks for clarifying that, really apprciate it! My knowledge of Hive on a deeper level is somewhat limited, this is why my idea seems to be lacking (especially when it comes to abuse). This really seems to be a tough nut to crack 😅 I hope we figure out a way eventually as we still need to make some leaps forward for broader adoption.
How is this different than one account being followed by hundreds of "inactive" brain-dead (removing the B in PoB) accounts on hive.vote? They have a ton of artificial weight behind their votes skewing the numbers.
One person saw your post and 200 people didn't see it, but voted. How is that a useful metric?
it's certainly different... but i'm also not saying the algorithms known as Hot and Trending are amazing at displaying the best content. I'd love to see better algorithms available.