The biggest objection to switching to 3 has been that library support hasn't been there. It's pretty clear that despite reservations about 3, the support is there now. Getting major libraries in the scientific community to focus attention at three can only be good for improvements in support, performance, and adoption.
Even sub-communites don't solve the problem. There are certain subreddits notorious for dogpiling on posts in other subreddits. The article mentions this indirectly by observing that what starts out on one platform can easily migrate across communities, even to communities you're not aware of.
There's a mismatch in scale: the number and diversity of members on platforms like Twitter is huge, but the ability to moderate is tiny. It would be interesting to see if a system could detect the dogpile effect, by looking at common phrases, checking if many users messaging one user all follow a common user, etc. and offer some sort of temporary relief until a controversy dies down. On the other hand, that's not actually solving the root of the problem.
HgFlow already exists and is actively developed: https://bitbucket.org/yujiewu/hgflow/wiki/Home. It even goes above and beyond GitFlow and adds generalized streams and sub-streams. I've never had to use those extra features, but the core model works well.
> People walk away from a game of Cards Against Humanity thinking that they’re hilarious.
Which people? My friends and I walk away from a game of Cards Against Humanity thinking we had a fun time coming up with stupid combinations for an hour. I can also turn the author's previous sentence around and state that there are probably many people that get too invested in elves and orcs as well.
> ... this passes once you’re more familiar with the language and are using it for real things
> ... a large enough subset of the community never reaches that stage
This is true about pretty much every programming language (and probably generalizable to most non-programming things too). The author just happened to apply it to Ruby and Haskell for some reason.
If there's a message such as "programmers should be more aware of the language's strengths and weaknesses and shouldn't feel overconfident", it's lost in the paragraphs of vague complaints.
Yes, if reusable/sharable queries are something you want, then you don't really have a choice but to expose an API for creating and retrieving them. Not sure if you're trying to make some other point related to the `Accept:` header the parent comment was about.
A fair point. I was referring to the api being separate from the frontend, not what the api is returning. But I could see someone designing their application the way you describe.
> A GET query is always a URI. Anyone can link to it.
Except this isn't really useful outside of bookmarks in a browser. Who would bookmark an API endpoint returning JSON? In code, it's just as easy to make a request with query parameters as it is with a post body.
We recognize that it's gotten a bit out of control lately, hence the topic during the meeting. I think we ended up with some good guidelines for managing it.
Then edit your question to acknowledge the "duplicate" and clearly explain why your question is different. It will automatically go into a Reopen Queue where users will evaluate it and vote to reopen. The system's not perfect, so if it still doesn't get reopened, open a discussion on meta.
Most comments should end up as an edit to the question or answer, clarifying some point, at which point they can be removed. The Question and Answer are the important things on Stack Overflow, not the communication that went into creating and tuning them.
The remaining comments are just fluff, such as "thanks", which can more appropriately (for the site) be expressed as an upvote or accept. Or they're asking a new question, in which case the parent's comment applies.
HN discussion: https://news.ycombinator.com/item?id=5204967