Trello – Google Doc style comments? Direct a Comment to a particular user

I believe that it's not automatic and that there are a group of employees who moderate these.

You could find your comment by using binary search and changing the url manually:

1. Start by getting the time of your comment from the activity log. Then go to the post and press on the time-and-date-text below one of the comments. The URL you go to should look something like this:

   https://www.facebook.com/{page}/posts/{postID}?comment_id={comment_id}&offset=0&total_comments=1000000

2. You are looking the offset for your comment, you know it should be between 0 (lower bound) and the number of comments (upper bound, In this case 1 000 000). Offset 0 means the newest post and 1000000 is the oldest post (in this case).

3. Let L denote the lower bound of offset, U denote the upper bound and M half way between L and U. (M = (L + U)/2)

4. Remove the comment_id from the URL and change the offset to M. The URL should look like this

   https://www.facebook.com/{page}/posts/{postID}?offset=M&total_comments=1000000

5. Look at the time on the comment there and compare it to the time of your post.

   • If your post is older, then the offset is to low. Let U = M and calculate a new M.
   • If your post is newer, then the offset is to high. Let L = M and calculate a new M.
   • If your post is from almost the same time, just press "Show more" until you find yours.

6. Go to step 4 and repeat until you found your comment.

This should give you the your comment within log2(1000000)=17 iterations, so even though it is complicated, you should find the comment a lot faster.

To find your comment (even) faster you could use the time difference to guess something between L and U other then M.

Optional:

Let t_L denote the time difference (e.g. in minutes) between post L and post U, and t_Y denote the time difference between your post and post U. Then the following formula for M should find your post in very few iterations (1, 2, 3 depending on how evenly the comments are spread):

M = (U - L) / t_L - L

But it probably won't save you much time because it is complicated to calculate the time difference in minutes.

If there is not a million comments you could probably skip the binary search and just guess from the time difference between the first comment, the last comment and your comment instead.