[wildwildtest]

I think the real lesson is that popularity is a poor way of understanding whether something is actually useful or good, it can indicate those qualities but meaningless without also applying our reasoning faculties too. The author appears to remain oblivious to any of the very good reasons why people don’t like alternative json formats and incapable of incorporating that into their understanding of people’s reaction. Citing market mechanics as a justification of “popularity = value” is blowing past the many examples of where free markets produce terrible approximations of value.

Also, I think it’s very poor character to say on one hand that people’s reaction doesn’t bother them and dedicate a third to a half of the article to Mitchell’s criticism specifically. I think Mitchell’s criticism is completely reasonable (I remember this as an era of “solving json” through multiple alternative formats) and drawing attention to what seems in retrospect as a mean-spirited act is actually quite a cynical attempt to strip the context of it and deliberately cast Mitchell in a poor light. Mitchell’s intention is very clearly a technically minded one to nudge people away from trying to massage long term, durable standards in often trivial, subjective ways with an end goal that can only result in format thrashing that doesn’t ever address anything substantial.

[mikkergp]

Except Mitchell explicitly called the author out in his satire and the author was right that Mitchell thought in retrospect it was a mistake to do so. Why is it in poor character if he’s expressing a view Mitchell agrees with.

Additionally in terms of the criticism, I observe the inverse phenomena to what you describe. You say that popularity isn’t a good indicator of goodness, but I think the inverse is goodness isn’t an indicator of eventual popularity. Some very elegant solutions just aren’t flexible enough for the real world. It is short sighted to try to argue some objective definition of badness because it doesn’t solve a problem you have or doesn’t fit your version of what a product like that should be. Sometimes in the absence of perfection people just have to get work done and want tools that make that easier.

[wildwildtest]

> Except …

I think that’s the sleight of hand occurring right, it’s reasonable to say he might mot have linked to it, but the purpose of the repo was not to shame the author. It might be my own predilection, but it strikes me as dishonest to say “I didn’t mind this” and then imply the opposite.

> Additionally …

Right, me too, they both occur. I don’t think I actually specified an argument as to why json5 is bad or even argued that it was. My comment is mainly about the article itself.

This whole paragraph reads as an argument to not use json5 given that it is very much trying to perfect the imperfect json, which most people just get on with.