To-do list before going public - TeX

Raaja

@marmot Finally I got a clear explanation for Ja mei ;)

user 3.14159 replying to Raaja

Ja mei.

Raaja

@marmot I just flagged your answer #2 to see how flag works, and unflagged it ofcourse

frougon

Thanks @JackDouglas, that's good to hear!

Jack Douglas

::: quote 543 15058 97165e8d4eb871ce5c5c85e6fa72da92515a45fb9260aa0c984157b0783e52f8 219,219,219 131,133,133
hackerb9 *— 6 days ago* *in [Can/should we start a Linux (or *nix) community on TopAnswers](/meta?room=543#c15058)*  
>Hi! I'd like to join the Beta test for the Unix site. I prefer my licenses to be upwards compatible with the future, so I always specify the version as an inequality (e.g., GPL ≥3.0). Any chance we can do that here? 
:::

Jack Douglas

@frougon thanks, you'll be pleased to hear we have already started work on this after it was requested earlier:

user 3.14159 replying to Jack Douglas

;-)

Jack Douglas replying to user 3.14159

Oh my!

user 3.14159 replying to Jack Douglas

It is impossible to really explain this. Some web sites claim that the translation is "Oh my" but this is not even close to the truth. "Ja mei" can have many different meanings, depending on the context and who says it. This is why it is so powerful.

Jack Douglas

@marmot is this a "Ja Mei" emoji, or am I not getting it: 🤭

Caleb replying to Jack Douglas

Maybe, but there are others with better expertise there. I don't expect it to be a hard thing.

samcarter replying to Jack Douglas

In case nobody here has experience with this, it might be worth a try to ask ArtOfCode over at codidact.

Jack Douglas

if not I guess perhaps it's the kind of self-contained task that someone else may take on…

Jack Douglas replying to Caleb

is this the sort of thing you'd be willing/able to help with once we have the source up?

Caleb

If you did a manual process asking people to verify with TA:UUID in their SE profiles you would actually be providing _more_ information to SE I believe.

Caleb replying to Jack Douglas

As far as I know, not a gram more than is already being manually entered into your database already in the form of a known ID for tha person on another site. At the end of the workflow all you get as a confirmation that your user is also user X on the oauth provider. The other side also gets a bit of data: that the user authorized themselves as something, but it doesn't establish the link the other way, only that they _used_ the workflow and that the redirect URL was site Y.

Jack Douglas replying to Caleb

how much privacy data do we leak if we go that route? I really know nothing at all about oauth

CarLaTeX replying to Skillmon

There is one, but as you said, it doesn't matter

Skillmon replying to CarLaTeX

If we're talking about the same site and the same 3 names, I did not see any of them. But honestly I don't really care, signing that or not doesn't change anything.

CarLaTeX replying to samcarter

Yes, there is the least seen on the site

samcarter replying to CarLaTeX

Really? I did not see any of the 3

CarLaTeX replying to samcarter

Not "all", there is one, curiosly

Raaja replying to Jack Douglas

I proposed this mainly due to a recent experience (may be others would have experienced it too). For instance, a user can abuse someone, and while trying to flag it, they may delete the entire post itself. There the moderation will not be effective as much as flagging a user directly. If you still think otherwise, then I have no problem. You can skip that too.

Raaja replying to Jack Douglas

ok I agree with your rationale

samcarter replying to CarLaTeX

Yeah, very sad to see them all missing from the list

CarLaTeX replying to samcarter

I didn't know the matter went even on press! OMG (please note who signed and who didn't)

samcarter

A nicely working example of a site using SO oauth: https://dearstackexchange.com/

Caleb

I assumed they didn't have one, but they do.

Caleb replying to Jack Douglas

Actually no. SO has an oauth endpoint, it should be quite easy to just authenticate and link the accounts for real. The workflow for that is actually pretty simple.

Jack Douglas replying to samcarter

good point but at least then only some will have to be done manually rather than all!

samcarter

(not every user on stackexchange is able to edit their profile, e.g. suspension, at least not without some tricks)

Caleb replying to user 3.14159

Jar jar binks.

Caleb replying to Jack Douglas

Yes that would do the trick. Maybe adding a blurb to the profile showing them what to paste, then a way to request verification. Later an automation that scraped the profile using XPath would be an easy addition to complete the loop.

user 3.14159 replying to Caleb

Ja mei. ;-)

Jack Douglas replying to Caleb

how about asking users to add their topanswers user id to their SE profile text (subtly, like "TA:123")?

Caleb

@marmot I don't even know what "Ja mei" is. Can it really be a blocker to make this site public if ⓐ it's something in the genre of emoji and ⓑ some/many people aren't even going to know what it is?

Caleb

Even if it turns into just a documented workflow and form for requesting the link that prompts them to add something to their profile like their UUID to make manual validation easy **and reliable**, that would be okay. I'm less worried about the manual work involved (that's for the admins to decide if they can handle) than the inherent insecurity of an honor system relying on a user here pinging to say they are X somewhere else. Even if the usernames and profile pictures look like a match those are all easy to impersonate, and PII isn't available on either end to check.

Caleb replying to samcarter

Yes, I proposed it as that. If it's not something we're going to get I would reconsider it being a hard requirement, but I don't think word should get around about an alternative without that being an option for those that want it. It doesn't seem fair without.

samcarter

@Caleb Do you consider validating SE identities a go/no-go point?  Personally I think its nice to have, but can come later.

samcarter replying to Caleb

Ok, I'll start going through the chat history and flagging the ones agreed there

Caleb

Hey @All can some of the rest of you flag answers here that we all (or so ~3 of us or so) have agreed are just general requests, not requirements to leave private beta. It would be a lot easier to see what the status is if we _only_ had blocking issues listed here.

Jack Douglas

and re your "An option to report users, similar to flags for questions" answer, I think we don't want that. The emphasis should be on the content — a bad user is one who is producing bad or harmful content, and then the content can be flagged. Of course we need systems that try and prevent bad content if multiple flags are raised for the same user.

Jack Douglas

@Raaja re your answer "Option to report an answer as a duplicate or in violation of academic-honesty", you have the option of flagging but also commenting with the specific issue with the post in the question chat room. Of course this is 100% public rather than private communication with the cleanup crew, but that is in keeping with the philosophy here of not having more than we need to behind closed doors. Although secrecy is intended to reduce drama I think it often has the opposite effect on SE.

Jack Douglas replying to samcarter

that is correct, normal flags hide a post from unregistered users, and then a further flag from a member of the 'post cleanup crew' effectively deletes the post.

CarLaTeX replying to samcarter

@Caleb and @Jack don't like it, so I prefer to delete it

samcarter replying to CarLaTeX

If I understand it correctly, flagging will hide the post from public view (not logged-in users) and then users with some kind of delete privilege are able to delete the post. (As for the imported question: you can now simply reimport it and the answers will be added - I tried this with the \pi numbering question and it worked)

CarLaTeX replying to samcarter

Flagging means we can delete useless posts (for example the one I imported without answers by mistake)?

samcarter

@raaja @carlatex @skillmon @marmot @caleb Now that flagging posts got implemented, is this sufficient to go public beta?

Raaja replying to CarLaTeX

^^

CarLaTeX replying to Raaja

I totally agree with you!

8 Answers