Search Ideas

I implemented this a while back.

X caches link previews, so old previews remain the same. But new previews feature the discussion title, see eg https://www.opengraph.xyz/url/https%3A%2F%2Fveritula.com%2Fdiscussions%2Fcriticisms-of-zcash.

Feature idea: pay people to criticize an idea.

You start a ‘criticism bounty’ of 100 bucks, say, which is prorated among eligible critics after some deadline.

The amount should be arbitrarily customizable (while covering transaction costs). Minimum of $5.

There could then be a page for bounties at /bounties. And a page listing a user’s bounties at /:username/bounties.

When starting a bounty, the user indicates terms such as what kinds of criticism they want. This way, they avoid having to pay people pointing out typos, say.

Anyone can start a bounty on any idea. There can only be one bounty per idea at a time.

To ensure a criticism is worthy of the bounty, the initiator gets a grace period of 24 hours at the end to review pending criticisms. Inaction automatically awards the bounty to all pending criticisms at the end of the grace period.

Feature idea: pay people to criticize your idea.

You start a ‘criticism bounty’ of 100 bucks, say, which is prorated among eligible critics after some deadline.

The amount should be arbitrarily customizable (while covering transaction costs). Minimum of $5.

There could then be a page for bounties at /bounties. And a page listing a user’s bounties at /:username/bounties.

When starting a bounty, the user indicates terms such as what kinds of criticism they want. This way, they avoid having to pay people pointing out typos, say.

Anyone can start a bounty on any idea. There can only be one bounty per idea at a time.

To ensure a criticism is worthy of the bounty, the initiator gets a grace period of 24 hours at the end to review pending criticisms. Inaction automatically awards the bounty to all pending criticisms at the end of the grace period.

No need, see #3420.

This is now a feature, see the ‘Funding’ section of a bounty.

When I revise a criticism, I can’t see what it criticises. The edit screen should show the parent idea, similar to when I write a new criticism.

Archiving covers this.

Could this feature be unified with #2811 somehow?

Bounties should be clear about what currency they are being paid out in.

When an idea has nothing but a code block, there’s too much of a margin at the bottom, between the block and the border of the highlight.

This issue didn’t only occur on ideas#show.

Fixed as of a44c6c0.

There’s still an issue on ideas#show. When an idea has nothing but a code block, there’s too much of a margin at the bottom, between the block and the border of the highlight.

When the code overflows horizontally, a subtle inset shadow on the side shows that you can scroll:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

const posts = [
  {id: 1, title: "Understanding JavaScript Closures in Depth", url: "https://example.com/articles/javascript-closures-deep-dive"},
  {id: 2, title: "A Complete Guide to Modern Web Development Practices", url: "https://example.com/articles/modern-web-dev-guide"},
  {id: 3, title: "Exploring the Node.js Event Loop and Async Patterns", url: "https://example.com/articles/nodejs-event-loop"}
];

function formatPost(post) {
  return `${post.id}: ${post.title} -> ${post.url}`;
}

function prettyPrint(posts) {
  return posts.map(formatPost).join(" | ");
}

console.log(prettyPrint(posts));

Done as of cc1ab95.

Ruby example:

1
2
3
4
5
6
7
8
9
10
11

def criticized? idea
  pending_criticisms(idea).any?
end

def pending_criticisms idea
  criticisms(idea).filter { |c| pending_criticisms(c).none? }
end

def criticisms idea
  children(idea).filter(&:criticism?)
end

JS example (h/t ChatGPT):

1
2
3
4
5
6
7
8
9
10
11

function criticized(idea) {
  return pendingCriticisms(idea).length > 0;
}

function pendingCriticisms(idea) {
  return criticisms(idea).filter(c => pendingCriticisms(c).length === 0);
}

function criticisms(idea) {
  return children(idea).filter(c => c.isCriticism);
}

Code blocks need syntax highlighting.

Veritula used to have this feature but I removed it when diffing changed.

Been trying a slight modification of bounties in prod for a couple of weeks or so. Working well so far.

@dirk-meulenbelt recently offered to chip in for a bounty I want to run. That got me thinking: multiple people should be able to fund bounties.

Tyler says:

No preview necessarily, or the first sentence upon mouse-over could work. I’m imagining a structural view independent of the main view. (Though still suggest looking at columns for each idea in the main view).

How would you preview text in nodes?

@tyler-mills says:

… I’m finding the threads a bit cumbersome to keep track of. Would love an option to have each top level idea in a column, and horizontal scrolling would be fine with me if there are many of them.

@tyler-mills says:

I keep coming back to a graph-based presentation. Every comment a node, edges red if ending in criticisms. I crave a way to see structurally how many red criticism threads and grey comment threads are stemming from a given idea. The red ones could be bold and bright if they lead to an uncriticized idea, else dim and thin. Then we can see at a glance which ideas are sources of more criticisms, and/or hold greater opportunities for further criticism — can see which ideas are “deeper” niches, one might say (..!). Have greater evolvability…

Basically not doable for the user with the current bubble+hashtag method. But again it could just be an optional view. I think I mentioned I find that Kialo does a cool job with their sun dial diagrams (which are optional).

In a future iteration, the user could additionally set a per-criticism ceiling. Which the site would recommend setting when using permissive terms.

This way, the user could set a total budget of $200, say, while capping each criticism at $30, for example. The first 6 eligible criticisms would each get $30, and the next one would get $20. The remaining criticisms would get nothing.

This approach effectively merges #3474 and #3472, giving users maximum flexibility to choose the best outcome depending on what kinds of criticism they anticipate getting based on their terms.

Let’s say somebody starts a bounty with permissive terms, asking for virtually any kind of criticism. They set a high ceiling, hoping for many submissions. $200, say.

If they only end up getting one or two small criticisms, for typos, say, they won’t like having to pay 100 bucks a pop.

In other words, the few criticisms you end up getting may not be worth the ceiling.

This approach is more complex for the bounty initiator than just indicating a total amount they are willing to spend (#3474). It’s best not to require users to do math.

Given the need for a deadline, all critics get paid at the same time anyway.