-15

I'm wondering about the reliability of the feature that allows to address other users in comments using the @user syntax.

I see the following problem:

User names can be changed. Because they are not unique the system isn't able to display the comments properly after a change since the system doesn't know which user was addressed once there are more users with that name. Result: The @user expression keeps unchanged, the name of the addressed user is now wrong.

Are there any plans to implement that feature in a reliable way? Or don't you care about that because you think it isn't a major problem?

I'm asking because today it happened the first time that I experienced the problem right in front of my eyes. A user who were addressed in a comment in the morning changed his name and now the comments point to somebody.

(It's this post on stackoverflow. The user who renamed itself was the OP)

  • 1
    Wrong. Only one user will be notified, being the latest Jon to post a comment. – ShaWiz May 5 '15 at 17:44
  • I'm almost 100% sure, that I've addressed users below a post where they didn't commented already. Are you 100% sure that I'm wrong? :) – hek2mgl May 5 '15 at 17:45
  • I can prove by bringing my sock who also starts with "Sha" and you can ping @Sha and I'll prove that only one got notification. :) – ShaWiz May 5 '15 at 17:45
  • You most likely talk about the case when you address the post author - they are always notified, even if they did not post comment before. – ShaWiz May 5 '15 at 17:46
  • First, thanks for your reply! Are you a developer at Stackexchange? I trust code more than 1000 words! :) – hek2mgl May 5 '15 at 17:46
  • No, no. Don't care about the addressed user being the OP too much. It was just the case in that example. It doesn't matter. – hek2mgl May 5 '15 at 17:48
  • @hek2mgl There are certain situations where you can @ reply to someone who hasn't commented, yes, but it's still a very limited subset of people. You can't just @ reply anyone. See the faq post for all of the details. – Servy May 5 '15 at 17:56
  • Three downs! Damn! I don't care about, but hey guys! I spent 30 minutes using Stackexchange's crap search engine to check if this is a duplicate and than other xx minutes to phrase my question. Still don't care, but common! @Servy Thanks for that link, let's see if it really answers my question. At least I put more effort in it than the guy who got 300+ ups. – hek2mgl May 5 '15 at 18:01
  • 2
    @hek2mgl You're asking to have a bunch of changes to a feature that doesn't have any of the problems you're claiming it has, what else would you expect to happen? – Servy May 5 '15 at 18:06
  • I admitted - in my question - that my first problem might not be a problem at all (the help doesn't state that), but the second still exists. Or do you like to follow a conversion where @foo is addressed - a user which doesn't exist anymore since he now named itself @bar. What if I wan't to address the author of that comment? Which should I use? I only have @foo -> not working. Not a problem? Downvotes for that? I don't get that! Do you like people who care about the site(s) or not? Do you expect me posting more stuff here? – hek2mgl May 5 '15 at 18:09
  • Also I tagged it explicitly as discussion, not feature request! Btw, I've never seen a domain before where user names aren't unique! – hek2mgl May 5 '15 at 18:12
  • @hek2mgl not really needed. Also that language isn't needed. The second point? Not that often tbh. – Tim May 5 '15 at 18:14
  • I'm not sure what you mean with that language. I was always polite and tried to address my concerns. Looks like there is not much more to do for me here. Thanks for the discussion which I've requested. – hek2mgl May 5 '15 at 18:19
  • @hek2mgl user name uniqueness is not considered critical. We're not a social network, so we aren't supposed to care someone is. It is the quality of the content that is supposed to matter. – psubsee2003 May 5 '15 at 18:19
  • 2
    @hek2mgl Mistagging a feature request as another tag doesn't make it not a feature request. You're requesting a feature, and it will be responded to accordingly. – Servy May 5 '15 at 18:22
9

I'm wondering about the reliability of the feature that allows to address other users in comments using the @user syntax.

"Wondering" means you don't really know enough about it to propose it's wrong or could be improved.
You really should do some reading and once you know more about the system you will be able to identify problems and think of possible improvements :)

Here's a good starting point:
How do comment @replies work?

The Stack sites are made by people with experience, and tried and tested functionality in place.

User names can be changed. Because they are not unique the system isn't able to display the comments properly after a change since the system doesn't know which user was addressed once there are more users with that name. Result: The @user expression keeps unchanged, the name of the addressed user is now wrong.

Yes this is true, you have identified a real potential issue, but it's not really a major or frequent issue.

The Questions and Answers are the most important things, comments are "useful" but not essential.
And going as far as being able to ping someone with @user in comments is nothing more than a nice feature to have, which is also "useful". They are not "essential" however.

I'm really not sure what else could be done, other than updating a HUGE load of databases/tables to update their names in comments. This would seem like a mammoth task for little gain. Certainly not worthwhile just so some of us can occasionally still ping their old comments.

Considering there are 4 million users on Stack Overflow (for example). That would mean scripts doing lots of searching through many DBs and tables to find all their comments in all questions they commented in, and then many DB updates to change their name. For all users who change their username.
This is very likely to have a detrimental effect on things such as site speed due to resources being used, and tables being locked.

Lastly, comments can be and often are cleaned up/deleted.

So changing usernames in comments is not really viable, or worthwhile.

Can you suggest a reason why it's a serious issue not being able to ping someone in comments because they have changed their name?
You can always ping the questioner and answerers, which is the important thing. Anything else is never going to be "important" to the site or the Q&A itself, and is always only "useful".

Are there any plans to implement that feature in a reliable way? Or don't you care about that because you think it isn't a major problem?

Most of what you raised isn't a problem at all. And/or there is probably no better ways to do it.

The site and users are open to suggestions, but you can't just jump on pointing out flaws in things you have not tried, tested, and understand fully. That's not fair really - and is probably the main reason for your downvotes.
You point out mistakes and problems where there is none, and suggest the site doesn't care, which is very inaccurate and again is an unfair "presumption" to make.

Have a read around the site about how comment notifications work, and the other suggestions people have made in an attempt to improve them.
After that, if you then have a query or good idea, don't hesitate to suggest/ask :)



Prior to the question edit:

if I have a conversation with Jon Doe on Stackoverflow and address him in a comment, multiple users might get noticed.

Again you state "might". How can you contest something you do not know about?
It's almost like you are saying "I do not know how this works, but if it does X when we do Y then that would be a problem".

The pings don't ping anyone/anywhere on the site. Otherwise that would just be silly.

It's only in the same thread where this happens.

See the link above, e.g.:

"Matching is performed in reverse chronological order, so if five people named John are participating, @john will match the most recent John. (Use the next rule to differentiate.)"

I don't see how else it could be done (without magic).

  • 1
    Thanks for your answer. I've edited my question but note that also in my original question I stated that I'm not sure about the first issue. My main concerns were about the second issue. – hek2mgl May 5 '15 at 18:35
  • @hek2mgl the second issue is quite rare to come across. Comments are secondary. Anything important should be edited in. – Tim May 5 '15 at 18:39
  • @Tim I admit that, for sure. I just thought that this is a place where we can discuss even about such minor issues. An answer like: "Hey, this isn't a big issue since ... " would be perfectly ok for me, I would had upvoted and accepted it. I was just wondering what others say. Probably this was the wrong place. Again, I have no issues with the current implementation, I was just wondering what will others say.... (Thanks for your reply! :) ) – hek2mgl May 5 '15 at 18:42
  • @hek2mgl right place. dvs here mean disagreement tho so. Also, there is a bias against low rep users, they are seen as unlikely to understand how the site works, and will be suggesting changes they don't understand :( me no like bias. – Tim May 5 '15 at 18:44
  • This is the right place to discuss, and to wonder what others say. That is the result you got ;) comments and my answer. But don't be dismayed from your downvotes. Even my answer got one without a comment as to why, and I don't see I stated anything wrong. Such is life on the old Stack Sea. – James May 5 '15 at 18:48
  • Yes I got some discussion, that's true! ;) But it was the wrong discussion and that's why I'm not satisfied. Basically I wanted to point out some issues with the fact that user names aren't unique on Stackexchange and discuss about pros and cons of uniqueness. Probably we could even find some ideas how to improve it. I'm a software developer, isn't this a valid question? I don't matter. I would appreciate if the discussion finally gets to the right direction. – hek2mgl May 5 '15 at 18:57
  • One of the issues you raised was not an issue at all. The other issue, while is an actual potential problem, is not really worth bothering about. I've edited my answer to follow your question edits :) – James May 5 '15 at 19:05
  • Thanks for your your efforts and your help. But the first paragraph, including the sentence The Stack sites are made by people with experience, and tried and tested functionality in place. . (cool!) ... isn't something I could accept. My issue is still open. – hek2mgl May 5 '15 at 19:09
  • Also I made nearly 10k comments on SO, this should be enough to talk about that feature, isn't it? Please respect that in your answer. – hek2mgl May 5 '15 at 19:11
  • 1
    @ShadowWizard2 Your having 10K comments in Stack Overflow is a valid argument against update your new username in comments - see my answer edit. For you alone that's a lot of DBs/tables to rake through and update. Someone could use SEDE to find out how many usernames are changed daily and total comments for all those users. That would give us an idea. (I don't use SEDE) – James May 5 '15 at 19:23
  • I don't expect a raspberry pi cluster driving the site. Also you admitted that those changes rarely happen. Such a change should be at least possible and feasible. – hek2mgl May 5 '15 at 22:15
  • I don't expect them to use 1U Dell 1950s either, but more resource usage means more servers and bandwidth. If something is worth doing I'm sure they would, but if it's largely pointless, then why waste resources or spend more money implementing it? There's dev time too, implementing and upkeep. – James May 5 '15 at 22:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .