Timeline for Allow high-rep users to recommend posts for and against tweeting
Current License: CC BY-SA 4.0
45 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 18, 2021 at 11:34 | history | edited | CommunityBot |
replaced http://chat.stackexchange.com with https://chat.stackexchange.com
|
|
| Apr 12, 2017 at 7:31 | history | edited | CommunityBot |
replaced http://programmers.stackexchange.com/ with https://softwareengineering.stackexchange.com/
|
|
| Mar 20, 2017 at 10:30 | history | edited | CommunityBot |
replaced http://meta.stackexchange.com/ with https://meta.stackexchange.com/
|
|
| Apr 24, 2014 at 14:01 | history | edited | CommunityBot |
Migration of MSO links to MSE links
|
|
| Jan 10, 2014 at 20:24 | answer | added | user194162 | timeline score: 3 | |
| Jan 10, 2014 at 19:31 | comment | added | Servy | @GlenH7 You're free to write up an answer if you want, I'm not sufficiently interested to write a detailed answer. | |
| Jan 10, 2014 at 19:17 | comment | added | user194162 | @Servy - I think you have a solid suggestion in filtering the high-rep users' votes to adjust the tweeting algorithm. Would you mind capturing those comments as an alternative answer? If you don't want to do so, would you mind if I put them into an answer? I think it would answer the intent of the question quite well along with some additional benefits such as keeping the UI simplified. | |
| Jan 10, 2014 at 19:05 | comment | added | gnat | @Servy that sounds reasonable to me. The question is, would it also sound reasonable to everyone at SE team? If yes, great, this way indeed looks more elegant to me. But if not... if someone raises and sticks to concern about voting anonymity... consider it buried | |
| Jan 10, 2014 at 19:01 | comment | added | Servy | @gnat Given that you wouldn't ever see the numeric value for this algorithm, it would likely be on questions with a fair number of votes already (thus providing a lot of data to mask the votes) etc. I can't see it being an issue. | |
| Jan 10, 2014 at 18:57 | history | edited | gnat | CC BY-SA 3.0 |
changed title to better fit question text
|
| Jan 10, 2014 at 18:47 | history | edited | gnat | CC BY-SA 3.0 |
The Whiteboard += http://chat.stackexchange.com/rooms/21/the-whiteboard "Programmers chat"
|
| Jan 10, 2014 at 18:41 | comment | added | gnat | @Servy with all due respect, I prefer to stick with "overloading" term - simply because I am not aware of any other use cases when up/down voting data was ever "filtered by rep" for whatever purposes. By the way, maybe this is because of concerns about keeping voting anonymous - if this is the case, "additional icons" approach could actually turn out a path of smaller resistance. Voting anonymity is a delicate matter | |
| Jan 10, 2014 at 18:34 | comment | added | Servy | @gnat It's not really overloading the buttons. The buttons simply provide votes. Those votes are used in all sorts of places for all sorts of things, ranging from rep changes, to post bans, to search results, to interesting questions, to readers, etc.. Lots of places use the votes on a post to do...something. This would be yet another "thing" consuming vote information. It's not changing the meaning of a vote; a vote still simply indicates helpfulness/usefulness. | |
| Jan 10, 2014 at 18:32 | comment | added | gnat | @Servy I see, thanks. That way looks worth considering. Overloading existing buttons to provide additional fine-tune "data stream", as a means to add the same functionality as requested but avoid UI complication, sounds neat | |
| Jan 10, 2014 at 18:29 | comment | added | Servy | @gnat And I'm saying that wherever you would otherwise use these new buttons, you can just use whatever those users used as their votes instead. It'll be the exact same information, it just involves zero UI changes instead of adding a whole new UI interaction to get information we already have. I fail to see why this is so complicated. The proposal is asking to collect information we already have. There's simply no reason to do that. | |
| Jan 10, 2014 at 18:20 | comment | added | gnat | @Servy my understanding is, it's supposed to "complement" what works now. Fine-tune it so to speak, as opposed to completely replacing it | |
| Jan 10, 2014 at 18:19 | comment | added | Servy | @gnat And how would that problem be solved by adding new buttons? You still don't have enough information in those cases either way. | |
| Jan 10, 2014 at 18:17 | comment | added | Jimmy Hoffa | @Servy in case you're unfamiliar with the background, you may like to have a quick look here which details more about the algo and has the official SE response; the precise technical problem they complained about was looking individually at the votes on questions/answers as opposed to aggregates. The high-rank-suggestion could be an aggregate perhaps making it feasible, I don't know, either way it might be good background for you to read. | |
| Jan 10, 2014 at 18:15 | comment | added | gnat | @Servy I see. That's an interesting alternative, but it may break at the opposite side of the issue I'm afraid. At sites / tags with small count of high rep users, this info won't make a reliable feed. Imagine like only 3-5 high rep users are online, they simply won't make stable representative "stream of data" for the feed | |
| Jan 10, 2014 at 18:10 | comment | added | Servy | @gnat I mean that the algorithm, instead of using some extra data field, can simply filter out votes from users without sufficient rep when determining what should be used. The information is all there; votes are only anonymized when displayed. It doesn't change how you vote, it's simply changing the algorithm to leverage the information already provided, namely the score of users above a given rep. | |
| Jan 10, 2014 at 18:03 | comment | added | gnat | @Servy you mean, voting buttons for these users could work as recommend / suspend tweeting? Functionally the same request, but instead of "two more icons", have up/down buttons be "overloaded" to additionally submit the tweeting feedback when clicked by hi-repz? | |
| Jan 10, 2014 at 18:01 | comment | added | Servy | @gnat You seem to have not been reading my comments. If you want to only look at high rep users then only look at high rep users, that's fine, but look at their votes, not an entirely separate button which they'll enter the exact same thing as their vote. | |
| Jan 10, 2014 at 17:58 | comment | added | gnat | @Servy information from votes is not the same, not even close. In hot questions, you get tens if not hundreds upvotes from careless passers by (aka "lemmings"), this completely obscures whatever votes could be there from "high rep users" (10-15-20Kers if I correctly understand this request). The point seems to be to "amplify" feedback of active site users which is currently lost in the noise | |
| Jan 10, 2014 at 17:54 | comment | added | Servy | @gnat Okay. I'm not opposed to it either; I'm just saying that we're already getting that information through votes, we simply need to use it. Getting the exact same information through a different button is just wasting people's time. | |
| Jan 10, 2014 at 17:49 | comment | added | gnat | @Servy "Providing a manual lever where trusted members of the community can endorse high-quality questions actually sounds like a good idea in principle" (Robert Harvey) | |
| Jan 10, 2014 at 17:37 | comment | added | Servy | @JimmyHoffa If it's completely impossible to change how up/downvotes are weighted why would it be easy to add an entirely new metric, with it's own weights, into the algorithm? You're claiming that no changes at all are realistic...except for this one? | |
| Jan 10, 2014 at 17:33 | comment | added | Jimmy Hoffa | @Servy I don't propose to know the technicals on the SE side of what is easier or harder. If someone from SE comes in and says this is far more onerous and not technically feasible, that is a completely sufficient response. It's worth asking though to see if it is possible because it would be a good benefit to people being able to advertise their sites more effectively. | |
| Jan 10, 2014 at 17:22 | comment | added | Servy | @JimmyHoffa And you think this is a less complex alternative? It's adding a ton more UI work and making the algorithm quite a bit more complex. How could it possibly be easier to implement? | |
| Jan 10, 2014 at 17:14 | comment | added | Jimmy Hoffa | @Servy That would be a happily accepted alternative, but due to the hotness formula algorithm improvements that have been suggested multiple times being rejected, this is an attempt to improve something without requesting the algorithm change as the sounds from SE are that the technicals of how the algorithms work makes changing them extremely onerous. | |
| Jan 10, 2014 at 17:13 | comment | added | Servy | @MichaelT So then just change the tweeting algorithm to only incorporate votes from high rep users, rather than presenting new buttons that are intended to be used identically to the voting buttons. | |
| Jan 10, 2014 at 17:09 | comment | added | user213963 | @Servy if this was something that was limited to people of a sufficiently high rep, one tends to be under the impression that they have a better idea of the site and how they want it presented to the populous at large. At worse, its no worse than it currently is. At best, we can work to avoid the belief that the twitter feed should be hooked to the close queue by presenting quality questions rather than popular meh questions. | |
| Jan 10, 2014 at 17:02 | comment | added | Servy | @MichaelT The fact remains that if people are going out of their way to upvote questions like this, and they're not being downvoted, then what makes you think there are going to be dramatically more votes to not tweet than votes to tweet? | |
| Jan 10, 2014 at 17:01 | answer | added | user102937 | timeline score: 2 | |
| Jan 10, 2014 at 16:57 | comment | added | user213963 | What this is asking is the ability to put on delay the tweet, let things that shouldn't get tweeted from getting tweeted, and allow the site to prepare (protecting the question, notices, deleting existing poor answers to avoid setting a poor example for others). Significant focused viewership can do significant damage to the site (mod flags, meta posts, "why did my answer get deleted", etc...). This is further increased when the mod count is lower on the site... (and we haven't been able to persuade Yannis to carry a pager hooked up to the twitter feed yet.) | |
| Jan 10, 2014 at 16:55 | comment | added | Servy | @MichaelT Yes, I'm not disputing that, I never was, I'm saying that this change wouldn't change that. It would be just as bad with this feature as without it in the situations you've described. | |
| Jan 10, 2014 at 16:54 | comment | added | gnat | @Servy per my reading, protect feature is proposed for "high rep users" - not to those who don't have the rep... | |
| Jan 10, 2014 at 16:53 | comment | added | user213963 | @Servy people who can and do downvote in an attempt to limit the hotness formula impacts are often hopelessly unable to have an effect once a question reaches a certain critical rep. Then it gets tweeted and the floodgates open and you get 5k views on a question in a day where the normal views on a question are ~500... and the rep explodes. Unless a mod is active enough to protect (10k can't protect within a time period) and put notices on it often gets a dozen of meh answers that ultimately leads to a poor impression for new visitors about the site's quality and experience. | |
| Jan 10, 2014 at 16:46 | comment | added | Servy | @MichaelT So you think those same people who upvote are going to not suggest tweeting, and those who don't have the rep and/or interest to downvote are going to go out of their way to suggest that it not be tweeted? | |
| Jan 10, 2014 at 16:41 | comment | added | user213963 | Having the questions hit the hot questions and tweet drastically increases the number of people who want to see it and upvote it on "interesting" criteria rather than the "useful" criteria. People who see it on the hot network and go "meh" to the question rarely visit the question, and those that do don't always have the rep to downvote if they were so inclined. | |
| Jan 10, 2014 at 16:41 | comment | added | Servy | @GlenH7 It sounds like such a question should still be downvoted then. And if the problem is nobody is doing that, then do you really think the people upvoting these questions are going to actively suggest they not be tweeted? These seem like the kinds of questions that, if people are going out of their way to upvote, that they'll go out of their way to tweet/suggest tweeting. | |
| Jan 10, 2014 at 16:38 | comment | added | user194162 | @Servy - this is a bit different. The questions are merely meh but not close worthy. And they could be tolerated, but you certainly don't want to advertise them for the site. The problem is that they'll gather up votes from the "popular / fun questions are great" camp. | |
| Jan 10, 2014 at 16:37 | comment | added | gnat | @Servy it's known for ages that one of these buttons (vote down) doesn't work in hot questions | |
| Jan 10, 2014 at 16:36 | answer | added | djechlin | timeline score: -2 | |
| Jan 10, 2014 at 16:32 | comment | added | Servy | We already have buttons for users to provide feedback on whether or not they feel a post is of high or low quality; it's up/down votes. | |
| Jan 10, 2014 at 16:29 | history | asked | user194162 | CC BY-SA 3.0 |