4 ms·

Ask HN: How can I make better comments?

Some of the comments on HN have been degrading slightly over my time on HN. I was wondering if there was a general method of how to create good an insightful comments on HN.

siong1987 18y ago
I found that the best way to submit a good comment is to edit your comment at least 3 times. Anyway, I found that this feature is a bit annoying especially some people edit their comments straight after someone replies to the comment. Sometimes, I even see some replies are totally unrelated to the comment because the commentor change the comment after people put replies to the comment. And, never write a reply that is against PG if you want to keep your karma point high. Some people will downvote your reply blindly.
- tokenadult 18y ago
  "never write a reply that is against PG if you want to keep your karma point high." Really? I'd expect pg himself, at least, to upvote a comment that gets him to think. I don't think I've tried yet the experiment of disagreeing with him directly. I'll observe the results if I do. To answer the main question, I like comments that grapple with the main point of the post or comment to which they are replying, and especially comments that link out to useful resources on the same or a related subject. A kind of comment I would like to see more of, a kind I have posted but don't get a lot of upvotes from, is asking a question along the lines of "How could we find out more about the facts of this issue?" or "What is the significance of this statement and its importance?" (You can see specific examples in the listing of my comments with low upvote totals. :) ) I really would like to know from the smart people here how they gather and evaluate evidence and how they make social policy trade-offs among competing interests.
  - jamesbritt 18y ago
    "Really? I'd expect pg himself, at least, to upvote a comment that gets him to think." Perhaps a better general version is, "Criticize the content of a comment, not the writer or their style." 'A kind of comment I would like to see more of, a kind I have posted but don't get a lot of upvotes from, is asking a question along the lines of "How could we find out more about the facts of this issue?" or "What is the significance of this statement and its importance?" ' From time to time I like to ask how someone happens to know whatever it is they are claiming. It does not seem tot be a popular question, though.
    - trapper 18y ago
      Agreed. I come from a medical background and we have to back everything with evidence. Even though it's computer science, most opinions of software guys tend to be backed up by blog posts or group think unfortunately. I am not sure how many people even read citeseer! Case in point: has there been any research touting the effectiveness of agile methods? What effect does pair programming have on productivity? Most opinions on the subject I have ever read haven't even looked, they just use anecdote and groupthink as their evidence. It's like listening to supplement pushers.
    - tokenadult 18y ago
      "From time to time I like to ask how someone happens to know whatever it is they are claiming. It does not seem tot be a popular question, though." Feel free to ask me that any time. I hope I notice your comments and upvote them when you do that.
gojomo 18y ago
Minimize. Don't say 'great article' or 'great comment'; use an upvote. Avoid cliches of phrase and thought, especially emotionally-charged cliches. Details from hard-won personal experience are always good. Armchair intuitions and snap judgments, rarely. Use the 'delay' setting, set for a few minutes, and reread your comment just as the delay is expiring. Fix typos and problems with clarity at any time, but avoid adding or removing substantive points without acknowledgment -- especially if there are replies. Read others' comments charitably. Think, "In what way could this make sense?", rather than "Aha, gotcha!" (Obsessively finding trivial bugs in code is a virtue; in conversation, it's rude.) Also, read articles charitably and considerate of their original context. A mood piece, a blog post, a popular-media account, etc. is written by an author with certain goals and for an imagined audience. To savage such an article because, once removed from its original context, it doesn't meet other standards or please some other alien audience is cheap, empty, artificial controversialism.

Some of the comments on HN have been degrading slightly over my time on HN. I was wondering if there was a general method of how to create good an insightful comments on HN.

siong1987 18y ago

I found that the best way to submit a good comment is to edit your comment at least 3 times. Anyway, I found that this feature is a bit annoying especially some people edit their comments straight after someone replies to the comment. Sometimes, I even see some replies are totally unrelated to the comment because the commentor change the comment after people put replies to the comment. And, never write a reply that is against PG if you want to keep your karma point high. Some people will downvote your reply blindly.

tokenadult 18y ago
"never write a reply that is against PG if you want to keep your karma point high." Really? I'd expect pg himself, at least, to upvote a comment that gets him to think. I don't think I've tried yet the experiment of disagreeing with him directly. I'll observe the results if I do. To answer the main question, I like comments that grapple with the main point of the post or comment to which they are replying, and especially comments that link out to useful resources on the same or a related subject. A kind of comment I would like to see more of, a kind I have posted but don't get a lot of upvotes from, is asking a question along the lines of "How could we find out more about the facts of this issue?" or "What is the significance of this statement and its importance?" (You can see specific examples in the listing of my comments with low upvote totals. :) ) I really would like to know from the smart people here how they gather and evaluate evidence and how they make social policy trade-offs among competing interests.
- jamesbritt 18y ago
  "Really? I'd expect pg himself, at least, to upvote a comment that gets him to think." Perhaps a better general version is, "Criticize the content of a comment, not the writer or their style." 'A kind of comment I would like to see more of, a kind I have posted but don't get a lot of upvotes from, is asking a question along the lines of "How could we find out more about the facts of this issue?" or "What is the significance of this statement and its importance?" ' From time to time I like to ask how someone happens to know whatever it is they are claiming. It does not seem tot be a popular question, though.
  - trapper 18y ago
    Agreed. I come from a medical background and we have to back everything with evidence. Even though it's computer science, most opinions of software guys tend to be backed up by blog posts or group think unfortunately. I am not sure how many people even read citeseer! Case in point: has there been any research touting the effectiveness of agile methods? What effect does pair programming have on productivity? Most opinions on the subject I have ever read haven't even looked, they just use anecdote and groupthink as their evidence. It's like listening to supplement pushers.
  - tokenadult 18y ago
    "From time to time I like to ask how someone happens to know whatever it is they are claiming. It does not seem tot be a popular question, though." Feel free to ask me that any time. I hope I notice your comments and upvote them when you do that.

gojomo 18y ago

Minimize. Don't say 'great article' or 'great comment'; use an upvote. Avoid cliches of phrase and thought, especially emotionally-charged cliches. Details from hard-won personal experience are always good. Armchair intuitions and snap judgments, rarely. Use the 'delay' setting, set for a few minutes, and reread your comment just as the delay is expiring. Fix typos and problems with clarity at any time, but avoid adding or removing substantive points without acknowledgment -- especially if there are replies. Read others' comments charitably. Think, "In what way could this make sense?", rather than "Aha, gotcha!" (Obsessively finding trivial bugs in code is a virtue; in conversation, it's rude.) Also, read articles charitably and considerate of their original context. A mood piece, a blog post, a popular-media account, etc. is written by an author with certain goals and for an imagined audience. To savage such an article because, once removed from its original context, it doesn't meet other standards or please some other alien audience is cheap, empty, artificial controversialism.