Re: Proposal: Nuke negative comments
[email protected] ("Larry Garfield")
| Newsgroups | php.doc |
|---|---|
| Message-ID | <[email protected]> |
On Sun, Jun 5, 2022, at 10:59 PM, Karoly Negyesi wrote: > Well, if you were looking for information on SNMP for Windows and you > can't find anything then you could still discover the note which > Christopher has indicated. > > It's not clear to me what we win by removing the downvoted ones. > > Having a hackathon which clears up curl_setopt would have, I'd argue, a > much bigger impact. Because, you know, there are an incredible number > of completely incorrect or just outdated (safe_mode, anyone?) comments > there -- massively upvoted, of course because they are old enough to > drink in any number of countries :P But I have not seen any > concentrated efforts on dealing with the comments but I might have > missed something. > > Drupal, for example, started a bug smash initiative (not my making nor > do I have anything to do with it) which is steadily trying to decrease > the cumulative bug age. > > Overall, both from a theoretical perspective and from a practical one, > deletion seems to be pointless to me. What does it enable which we > can't do currently and how does it help with things that we should do? There's a clear consensus that we should be migrating meaningful comments into the docs pages proper and deleting those that are not meaningful. I don't think that's disputed. However, there are currently maybe 5 people who work on that from time to time. If that. There are currently 23,862 comments in the system. We don't have a century or two to get through them all at the current rate. We need some kind of automation to help. A heuristic of "if it's downvoted below -X, odds are it's worthless and we can kill it" is imperfect, but it's at least some kind of automation to make that task easier by removing worthless floatsam. That can give us a more accurate picture of which pages have the most comments worth considering; right now we get a weekly update on the pages with the most comments, but that doesn't tell us where the most "comments worth integrating" are. Purging comments we are reasonably confident are crap will save us time, and give us a somewhat more accurate picture of where to put our collectively very limited time. It's a decidedly imperfect mechanism and heuristic, I agree, but so far no one has proposed a better one. Throwing more warm bodies at the problem would be nice, but we are currently lacking in warm bodies and they do need to be at least a little bit more skilled and versed than "has a pulse." --Larry Garfield