[tantek]aciccarello, how is your IWC SD post going? if it helps in sympathy, it has taken me a few partial-attention hours to write-up my post about the Gresham’s law effect of developer conversations vs users
Loqi[preview] [Tantek Çelik] Similar to @paulgraham.com (@paulg@mas.to @paulg)’s observation about trolls¹, there’s a sort of Gresham's Law of developers (vs users): developers are willing to use a forum with a lot of users in it, but users aren’t willing to use a forum w...
[KevinMarks]“The core users of News.YC are mostly refugees from other sites that were overrun by trolls. They feel about trolls roughly the way refugees from Cuba or Eastern Europe feel about dictatorships” is rather loaded
[KevinMarks]Yes, that's good. Less snarky than I would have been tempted to write. Gresham's law is an interesting analogy, as what it says is that people pass on the forged money and keep the real coins, so the forgeries circulate more rapidly. Which is rather like the way sensational lies spread faster than truth.
[tantek]thanks for that review. the core point (warning) of the essay stands IMO even if the solutions noted therein haven't worked, which is ok because I'm not referencing it for the solutions, more the problem statement
[tantek]ok reviews and ratings folks, where did the default 1-5 rating scale come from? I feel I have found numerous flaws that would be solved by a simple addition of the use of "0" (zero) to ground the lower end of the scale, or in many cases reducing the rating scale to -1, 0, 1. What am I missing? Why are people/experts and orgs so stuck on 1-5 (star) ratings? (yes I'm thinking about a blog post about this)
[tantek][KevinMarks] no harder than showing one star out of five, because presumably you need to show four empty stars, otherwise one star looks like AWESOME, this thing is a STAR! (like the old favorites on Twitter) ⭐️
gibruin the past, I've seen this problem solved by using five star outlines (0) and for each rating you simply show a full star e.g. five star outlines, three of them filled is a rating of 3 out of 5
[tantek][KevinMarks] good example. Also even being *mentioned* by Michelin, i.e. being on their published list, without any stars, is a positive evaluation.
mandarisI recently thought about my own rating system. I determined that if I was to give something zero 0️⃣ stars, it would indicate that it was so bad that I couldn’t finish it.
[snarfed]single personal data point, I found that -1/0/1 was the right level of granularity for my own ratings, eg https://snarfed.org/books . I ended up thinking too much about 0/1-5
[tantek]interestingly enough, Swarm/Foursquare uses a -1/0/1 scale for rating (verb) then displays a (0.0-10.0) rating (noun) as a way of doing some sort of fancy math(s)
[tantek]^ that kind of intimidation (or fear of reprisal from the venue) seems like it would produce skewed reviews and a good reason for doing Swarm-style friends-only and anonymous aggregated ratings
fury999io, [tantek], plantroon, bterry and a51 joined the channel
LoqiIt looks like we don't have a page for "publishing client" yet. Would you like to create it? (Or just say "publishing client is ____", a sentence describing the term)
LoqiIt looks like we don't have a page for "blogging client" yet. Would you like to create it? (Or just say "blogging client is ____", a sentence describing the term)
[Murray]Heh, I started with a 1-5 rating. Then I added 6. Then I added .5 stops. Then I started using 0. It seems to have settled there. For some things, I still use 1-5 (I guess 0-5, realistically), but for stuff like reviews a 0-6, with half-stop breaks, fits my own mental model better.
[Murray]But I do think that "fits my mental model" is the only _useful_ part of that. That works for me. I imagine it won't work for other people. Personally, -1, 0, 1 would be far too restrictive, and feel too harsh. That seems like there are two failing grades and one pass, seems pessimistic (though I understand that won't be the way other people read/think about it)
[Murray]I find it useful to have a halfway point that is both a whole number *and* considered positive. I'm okay with average; average should be an okay grade 😄