like a Hacker News Canon are good ideas that could be implemented much like searchyc.com, as totally separate projects.
I propose a solution, for consideration by the community:
"If you're posting to inform us of the existence of something, assume we already know."
Don't post a link to arduino.cc.
Maybe post a link to a fairly ordinary Arduino project.
Do post a link to an innovative new prototyping framework or exceptional new project.
Do post a link to yourown Arduino project.
Don't post a link to a programming language's home page.
Maybe post a link to an article about a language.
Do post a link to the home page of a newly released language.
Do post a link to yourown code or experience with the language.
We already know about rasterbator, Io, and Arduino. C'mon. We're hackers.
( anyway, I offer this as an idea for potential adoption by HN, not as a criticism; at the moment, a LOT of submissions seem to be of the "Whoa, I stumbled across a cool X!" class. That seems, at first blush, like a good recipe; in practice, I think it leads to more reddit-ish submissions than HN-ish, particularly as HN grows. )
To be fair, I've read a couple of articles like that and (to me, at least) they've been interesting because they'll sometimes explain the approach they had to take to support state in something as interactive as a game. So I always have a kind of morbid fascination with Haskell articles. It's so foreign; like a silicon-based life form. Still, my morbid fascinations will lead me to find those articles on my own, and do, before I see them linked on HN.
But if our goal is to "smelt an ever-purer" HN, it seems we should only submit a link to 'implementation of Canonical Application X in language L' if it answers one or both of the following:
1) How is the approach (non-trivially) different in L than in most languages? Did the language make something harder or easier? Why?
2) What about the implementer's experience is of value to us? Any gotchas to look out for?
