GitHub Deceased User Policy | Hacker News Reader

GitHub Deceased User Policy

docs.github.com

244 pointsbobbiechen105 comments

If one of your friends is a freenode user and comes down with a slight case of death, please contact the freenode staff via chat.freenode.net #freenode and their account will be memorialised. (I had to do one of these a few days ago and it's sadly not the first)

A lot of providers will be happy to figure something out, too - the previous-to-this one situation the deceased also had a linode account and once I got through to them they put the deceased's account onto a deep freeze where the systems were off but the data was kept to give the deceased's family time to get everything else done, with an understanding that once they sent a death certificate through that would allow for the data to be released to them.

Also, if you're lost, contact me as 'mst' on irc.freenode.net or irc.perl.org - I make no guarantees as to how much I can help but everything I've learned from dealing with my friends' passing I'll be happy to do my best to use to help you deal with your friend's passing. My friends would, if they were around to do so, appreciate the re-use of effort and sharing of knowledge.

Sorry for your loss, I wish you all the best

Thank you. I get better at the "paperwork" parts every time.

And I know my friends I've done it for before wouldn't be surprised that I keep volunteering to do it, and would, if they were around to do so, grin and say "yep, that's mst" at the fact I'm willing to use my experience on behalf of them to make it easier for others.

This really isn't a selfless offer. It makes -me- feel better, and I know they'd approve. So, yeah, seriously people, if you need a hand with this stuff, come poke me with a stick (and I mean that in general, freenode I know far too well, anywhere else I might not know that well but I'll try to help anyway because it'll be easier for me to think it through than it will for you if you just lost a friend).

> slight case of death

What counts as slight? Also, it sounds like something I would hear on an Oversimplified history video.

Mostly just my dark/gallows humour tbh.

Just like there's not really any such thing as 'slightly pregnant'.

en_UK understatement, because I've cried enough already so I prefer to remain as dry as possible from here on in.

It's been a while, but I think it's a Hitchhiker's Guide reference.

Seems an attempt to lighten the mood when discussing a serious topic. Surely you've seen this before.

For answers to this and other questions of life, the universe, and everything, stop whatever you are doing and download a copy of The Hitchhiker's Guide to the Galaxy.

For best results, get the first 5 books of the trilogy and do nothing until you have read them all. It will change your life. :)

What does memorialized mean in IRC?

Reserved nickname?

Basically the account gets marked memorialized and then will never be expired.

So, yeah, as the other reply said, "reserved nickname", but more precisely "reserved accountname indefinitely".

I find it a bit weird that you let nicknames expire for living people if you recognize the importance of keeping someone's identity constant.

Thank you for sharing, and please accept my condolences.

About memorialising IRC nicks, I understand why this is done but maybe it doesn't need to be a perpetual reservation?

When an amateur radio operator passes, their callsign becomes a SK (silent-key):

http://www.silentkeyhq.com/main.php

After a period of time, the callsign can be allocated to somebody else. Often this is a descendant of the operator, but not always. Usually the community is in fact happy to see a callsign remain active.

In some cases there are multi-generational callsigns that have hundred year histories behind them, which I think is a wonderful connection to maintain throughout our shared history, and future.

For usernames / callsigns there are other advantages to reserve them indefinitely like the possibility of them being impersonated. This is compounded because of the fact that on IRC your only method of communication is the internet.

I lost my fiancée to cancer a few years ago, and this just reminded me that I suggested she create a GitHub account years ago. I didn't know she had used it, but this post prompted me to check, and it turns out she created some (good) issues on a few Minecraft/CraftBukkit projects.

It's great to see companies dedicating resources to putting these policies in place. It can make a huge difference while someone is going through the most difficult time in their life.

Sad to hear about your fiancee. Good that this helps in keeping her alive in your memories.

Speaking of bukkit, bukkit based minecraft servers are still cool. There's a new one called Paper that's been working out well for my small group.

+1 for Paper, also works well with GeyserMC [0] so your Bedrock friends can join. BTW, the GeyserMC community is really great, top notch support via Discord with a fancy integrated bot and automated builds that you can just drop into you plugins folder.

I use this Docker image with Docker-compose: [1]

[0] https://geysermc.org/

[1] https://github.com/itzg/docker-minecraft-server

One thing to note is that paper breaks technical redstone builds. For redstone parity, Fabric [0] is what we use, along with Phosphor + Lithium + Carpet.

[0]: https://fabricmc.net/

I set up a Paper server for my kids. The modding ability is pretty cool, and easy to get into. I made a mod for them which rings a bell and announces the server time on the hour. They're less likely to lose track of time now. WorldGuard is cool because you can define regions and only allow certain players or groups of players permission to use that space. I can also prevent mobs from spawning in "town". Neat stuff.

I know that I said that I set it up for my kids, but I really did it for me as well!

GitHub/Microsoft is really setting themselves up to be a customer-centric (different to user-centric!) tech platform.

If you try doing this process with Google, you are given a canned form and you'll probably (not) hear back from a robot.

Yes, even LinkedIn is problematic. My dad is still listed on LinkedIn and Google while he passed away few years ago. Total silence from both, even when sending the death certificate.

Google has Inactive Account Manager, a dead man's switch for your account: https://support.google.com/accounts/answer/3036546?hl=en

Their activity detection does not inspire confidence. Google is an incredibly bad proxy for user activity.

Edit: downvoted but no substantive responses? Guidelines are for thee but not for mee, I guess.

a) The guidelines don't say anything about needing to give a substantive response instead of downvoting.

b) The guidelines do say not to complain about voting.

c) Both the creator of HN and its moderators have repeatedly said that downvoting for disagreement is perfectly fine.

"Downvoting has always been used to express disagreement." Paul Graham, February 16, 2008.

https://news.ycombinator.com/item?id=16131314

off topic, and it makes for boring conversation. but:

> "Downvoting has always been used to express disagreement." Paul Graham, February 16, 2008.

this is something that bothered me for years. I know this is how people behave, that the workflow allows and enables it, and the culture we've built around our tools enforces that. When I watch my own feelings before downvoting it isn't my rational self but my petty lizard-brain which does this. It feels good for 1 second before I feel small and petty. Often it bothers me that I did it and I return to the button 3 minutes later undoing the downvote for a neutral no-vote.

Who is served by the downvote? Is it the downvoter who feels upset another person disagreeing and now able to vent? Or the downvotee who can use the experience to learn what language to avoid? Or is it only the platform provider that benefits? If it's only the platform it seems like a very blunt and rudimentary tool IMHO. Perhaps a downvote is still less inflammatory for discourse than an angry reply and so a compromise? Regardless it's still lizard-brained and impulsive every time I use it and the more I think about it the more I'm puzzled that this is form of behavior modification is all we have after so many years on the Web. The biggest lie we tell ourselves is that sharing these thoughts are actually well thought-through "opinions" when in reality it's just "thinking out loud" or "reactions to other people thinking out loud". It's not like writing a letter, diary or journal. The Tech moves fast but our social collective behavior takes millenia to adjust and cope.

Anyway I wish there was a switch that allowed me to hide the downvote button, and if I wanted to enable it, then it could only be done by a forced delay of X hours into the future when activated. I'd like to eliminate downvoting as a behavior in myself in 2021 - not for others sake but for my own inner peace. And also because none of it matters really.

Edit: I took a quick look at this at it's even easier than that. You can do it with just stylus/Stylish.

  @-moz-document domain("news.ycombinator.com") {
  [title=downvote] {
      display: none
  }
  }

As a personal experiment, I'm going to leave this enabled for myself and see how long before I find the urge to downvote some comment. :D

Original: You could hide the button with TamperMonkey pretty easily (send an email if you genuinely want to, looked at it, and couldn’t figure it out, and I’ll try to help)

Then, the “only after reflection” I’d implement by forcing myself to use another browser or toggling the element title or the extension. I don’t downvote very often (probably 20:1 up to down vote ratio), but I do know that some of mine are still lizard brain.

thanks I will try your style receipt. it's fascinating to see how these things effects one emotionally.

I found a way here on HN to use uBlock to set Dark Mode and undo the downvote colors. I don't remember who suggested the Dark Mode. This is from "My Filters"

  ycombinator.com##html:style(background-color: #3e3e42 !important)
  ycombinator.com###hnmain:style(background-color: #2E2E31 !important)
  ycombinator.com##.age:style(color: #ccc !important)
  ycombinator.com##.c00:style(color: #ccc !important)
  ycombinator.com##.c5a:style(color: #ccc !important)
  ycombinator.com##.c73:style(color: #ccc !important)
  ycombinator.com##.c82:style(color: #ccc !important)
  ycombinator.com##.c88:style(color: #ccc !important)
  ycombinator.com##.c9c:style(color: #ccc !important)
  ycombinator.com##.cae:style(color: #ccc !important)
  ycombinator.com##.cbe:style(color: #ccc !important)
  ycombinator.com##.cce:style(color: #ccc !important)
  ycombinator.com##.cdd:style(color: #ccc !important)
  ycombinator.com##.comhead > a:style(color: #dedede !important)
  ycombinator.com##.comhead:style(color: #dedede !important)
  ycombinator.com##.hnuser:style(color: #ccc !important)
  ycombinator.com##.subtext > a:link:style(color: #ccc !important)
  ycombinator.com##.subtext > a:style(color: #ccc !important)
  ycombinator.com##.subtext:style(color: #ccc !important)
  ycombinator.com##.title > a:style(color: #dedede !important)
  ycombinator.com##a:link:style(color: #ccc !important)

The .c00 is the default text and the .c5a through .cdd are the 9 shades of gray.

> Who is served by the downvote?

I’m served by the downvote in knowing that what I said was downvoted. Usually, I don’t need a substantive conversation from everyone who dislikes my comment. The signal that someone dislikes is useful enough.

It’s like in a conversation, sometimes I’ll see a frown or drawn up eyes. I don’t want, or need, every time I say something stupid to have the other conversants stop me and say “well prepend, in that usage...”

I think there’s a mismatch of conversation goals. The school of thought that downvotes shouldn’t be used seems to value engagement over value and content. My goal is to learn and teach the most through comments. To convince and be convinced. As efficiently as possible.

Looking back at comments that are -n the downvotes are usually enough to understand what I did wrong and how to improve. Not always, but usually. And I would have hated it instead of that score I had n+2 comments engaging with me for how they didn’t like it. Or worse, no comment and no downvote. I think the latter is more likely as while I will downvote things I don’t like, I’ll rarely comment unless I have something substantive to say.

I think one of the main defects in Facebook, Twitter, etc is that there is no downvote, only upvote. So there’s no social norm to signify displeasure so we get into this weird cycle where people only know what works and it’s harder to know what doesn’t work.

I think social medial would be better off if we could distinguish the 1million upvoted content vs the 1million up, 1 million down, net zero content. But social media doesn’t really care about providing great content, they only want content good enough to keep me screaming and clicking.

the whole UX of down and upvoting is flawed from the onset, the binary choice of reactions is the real problem and is not how human interactions work

Platforms should provide more options for feedback on a given post other than up or down voting.

> Their activity detection does not inspire confidence. Google is an incredibly bad proxy for user activity.

I mostly agree with you on Google but your post has no substance either.

I mean, if you use Gmail for instance it's incredibly likely I'll be dead when I stop checking my email for a year.

Where in the guidelines does it say you must provide a comment if you're going to downvote? I've just read them twice and don't see it. In fact, the only thing I see relevant to your comment is:

> Please don't comment about the voting on comments. It never does any good, and it makes boring reading.

Thanks, I wasn't aware of this service, even though I've been a pretty heavy Google services user for (looks it up)...holy crap, almost 17 years now.

I think you are giving them too much credit - Microsoft won’t talk to anyone who isn’t a partner and GitHub tends to just be silent on anything they can’t cut and paste an answer to. It’s more an issue that mumbling on about your algorithm and submitting support tickets while bereaved kin desperately seek access to a loved one’s account is just incredibly bad PR and always makes the front page - and nobody wants the job of dealing with that.

Microsoft will definitely speak to people who aren’t partners - I work for a small business (12 people) and the two times I have had to go to Microsoft for support on 365 they have been incredibly helpful.

That's support you pay for tho. We also were surprised when i worked at a 9 person company to get someone from google on the line too when we were considering using google maps api for a project. Support which wasn't payed for but which I assumes comes from the expectation that we might become a paying customer.

The other side is that with some companies you can’t even pay for support if you wanted. I’m looking at you, Google.

Good luck with getting to Google support if you are paying customer.

"Terraria canceled on Stadia after developer gets locked out of his Google accounts" is just recent example

Have to agree with the other commenter here. I'm not a big Microsoft fan, but for organizations who use them that I've helped, I found it relatively straight-forward to get through and talk to someone at Microsoft when needed.

I disagree. I had an issue last week with the billing platform and contacted support. Was resolved in the same day. I was expecting that I won't get a response (they have millions of users?) but I did. And I'm not a big customer, or a partner or anything that close.

I can testify once you hit a tricky point with GitHub support they're completely useless. And when push comes to shove they'll start playing games - leaving off names in emails so it can't be traced back to a specific employee, sending response that look like they're internal comments, etc. Honestly it was a level of immaturity I didn't expect from GitHub. I went from someone who owned tshirts, hoodies, octopints, etc to someone that only uses them if my work forces me to.

That's not really true in my experience, I've had some issues with quotas being displayed incorrectly on my completely free account recently and got an answer within a few hours, saying that they reset the tracking and relayed a bug report to the team. I honestly think that for now, you can still depend on them to do their due diligence with support.

Anecdote: I have an Azure account with a monthly spend of around $100-200 (so, a very small customer). I ran into an issue with Active Directory B2C a couple of years back. I reached out - despite not having a support plan - and ended up with a Microsoft Engineer sharing my screen and helping me fix the issue.

On my local market it's the reason Google cloud adoption among medium to large companies and institutions is laughably low. Microsoft's support exists, it's a real argument towards them.

It may depend on the services. A few years ago I lost an old email account, a hotmail one, because it was flagged for being potentially compromised. An account flagged as so has its password wiped and you have to reset it by following a link that is sent to your recovery email address.

Somehow, and I don't know how my 10yo-self managed to do this, but the account's mail and the recovery mail were the same, so I wasn't able to recover my account this way.

The only alternative was to fill up a form with your full name, date of birth (probably lied about it at this time), passwords you may have used for this account, people to whom I had sent emails (I only used it to register to websites)...

I have never been able to do a "perfect score" to this test monitored only by robots. Now my account, from what I've read online, is probably disabled forever due to a 2 years inactivity.

So, on top of financial debt, you can also transfer technical debt to your successors ?

Immortality (in minds of pissed off relatives), here I come!

They say that every man has two deaths. The first when they breathe for the last time. The second when their name is screamed out and cursed at for the last time.

*the second when their last bug in production code is squashed

Sweet, I'm going to be immortal.

A former co-worker of mine passed away 8 or 9 years ago, and every year I get a LinkedIn notification asking me to congratulate them on their upcoming work anniversary. It's terrible.

Same thing was happening to me - LinkedIn has a solid deceased policy. I submitted a request with a link to a public obituary, and less than 24hrs later their page was gone. Bittersweet, I’ll miss seeing them pop up now and then, but also won’t feel like they are being disrespected by inaccurate online updates and notifications.

https://www.linkedin.com/help/linkedin/answer/2842/deceased-...

I too submitted a request to LI regarding the passing of a colleague. Support handled it the very same day.

Facebook sends me emails every year suggesting that I post birthday wishes to deceased relatives.

I was thinking the other day that I would like to open-source all my random projects or startup files if I die but I couldn't think of a good way to do it.

I guess maybe a private repo with instructions could do, but it requires a technical person in the family or friends to check it if/when GH gives them access. The way that Facebook does it with their memorial service is not bad. You can set what happens with the account and repos if a death cert is received.

As an option you can do dead man switch + some automation. Once per month script sends you an email with link. If you don't click the link in 1-2 months, script opens everything.

Of course, that won't handle corner cases like jail or coma, but should work for most everyday scenarios.

I am building such service, will release it in 2 weeks. It would check on you by email or sms in periods you define and if the trigger is met (eg you did not reply in X days/hours) it would send your predefined messages or webhooks or zapps or...

I think most complex part there is testing it from receiver's side.

Do you have a link to this? I would be interested in checking it out.

I have a document outlining to my (non-technical) wife where all the secret keys can be found, and some suggestions on friends to contact for help using them.

GitHub has an account successor system. The successor can be anyone. I set a technical friend of mine as my account successor.

https://docs.github.com/en/github/setting-up-and-managing-yo...

Why not open source them now?

(That's been my approach: open source everything by default)

Can it be done by api? If so, you can setup a service that will do it automatically after a year, and just set a reminder to rest the timer every 6 months or so.

Code isn't worth as much as people think without the humans.

I would challenge you to ask yourself why you want to wait until you're dead to share these things. Why not share some of them now?

This is great, it will be great if tech companies can also start working with the government systems which recognize the legal heirs as well. In the case of GitHub it may not be so pertinent, but with email ids etc. where bank account details and OTPs are sent to, it may be important for email providers to recognize real world legal entities and work with the usual estate management legislations.

This is not a new problem. This happens all the time when someone dies: Executors must put the deceased' affairs in order and contact all the companies the deceased had business with.

It's just that tech companies tend to think that this does not apply to them and that they don't need to be able to deal with actual mail or this sort of process.

Github's policy is simply how things ought to be (and credit to them for doing it). People find it so great simply because the industry is so appallingly bad.

I guess somewhat related to this: "Respectfully dealing with fedizens who are deceased" [0].

[0]: https://socialhub.activitypub.rocks/t/respectfully-dealing-w...

Is there a standard internet graveyard? For example, some place where all of your social media, blog posts, code and platform data can live under one ID after you pass away?

Interesting idea. Like a centralized memorial with curated data from your online presence over time?

Pretty much. I’ve always heard people say that they want places like Twitter and IG to keep accounts open for memories, but what you described is exactly it; a centralized memorial.

Everyone is acting like this is a great policy, except even if you specify a successor it only lets them transfer public repositories.

Worth mentioning, this wasn't always the case. Github didn't really have much of a policy until recently, if memory serves.

The author of the Sprout library, Bolero Murakami, passed away a while back and I took it upon myself to ask Github what they do in these circumstances. It was shortly after that IIRC that they introduced the account successor settings functionality and updated their death policy.

This part of the internet's functioning has always fascinated me, even since a kid - the fact that your online presence acts as sort of a living obituary, and what happens to your data after you lose ultimate control over it.

The website is not reachable anymore looks like. For me it just keeps loading.

> Error 503 first byte timeout

> first byte timeout

> Guru Mediation:

> Details: cache-hhn11525-HHN 1613548903 1064572670

> Varnish cache server

It has deceased.

  fn main() {
      println!("{}", "GitHub Deceased User Policy has deceased.\n".repeat(2021));
  }

Edit for the downvoters:

- https://play.rust-lang.org/?version=stable&mode=debug&editio...

- https://gist.github.com/rust-play/9979b95e1f4cb81e75a0dc72af...

Pining for the fjords

Their entire docs site is down, yep.

https://www.githubstatus.com/incidents/yxpgm9766lp0

Sometime ago I stumpled upon a github repo that interested me. Due to several reasons I suspect the maintainer might be dead. I tried to verify it somehow but no luck. Might be easier if you are a Greek citizen. Suggestion? Of-course a persons death is tragic, but it is also a shame no one will start maintaining a fork while the original repo is in limbo...

I feel like if its clear the maintainer has dropped off the internet, forking and dropping an issue "Something is up with XYZ, lets work in this forked repo until its clear whats going on" is fine. If they come back you can still offer to shut down your fork/help them transfer added content/...

It’s nice that they built something into the service but I think can this problem can be solved independent of the service.

One of my weekend project ideas is a small deadman-switch service that pings you every X days to press a button, if it hasn’t heard from you in Y days it sends a private key encrypted file and some instructions. Think onetimesecret with a button and a timer.

https://www.deadmansswitch.net/

Built by /user?id=StravosK

Now the question is: What happens to the DeadMansSwitch service if something were to happen to its maker? It's like we need to form mini multi-generational cults to keep human services like that going for as long as humans keep going until the point they stop.

Uhhh...you mean like, a company?

Yeah, but multi-generational cult sounds way cooler

The research people have been doing into conveying the dangers of nuclear dumps to potentially primitive future peoples is quite fascinating.

Agreed. https://en.m.wikipedia.org/wiki/Long-time_nuclear_waste_warn...

In futureLang:

- Hey, we found some hieroglyphs, in better shape than the ones near the pyramids, do you recognise the gods?

- No, but they all seem to say similar things: Don't open or "Calamities" or something. I wonder what their primitives beliefs were based on... Let's dig!

Maybe I should directly apply that language to my post-mortem open source projects:

This repository is not a place of honor. No highly esteemed code is placed here. Nothing valued is here. What is here is dangerous and repulsive to us.

This sounds like an interesting idea! I’m eager to help any way I can.

What have we done?

https://www.githubstatus.com/incidents/yxpgm9766lp0

It's thoughtful they have a policy at all. I'm sure, at their scale, this happens all the time. I just can't stop wondering how much of a vector for abuse this is? Ie. hostile takeovers of accounts. Yeah, I get it, they probably have some checks in place, but I imagine the legalities are very different per jurisdiction, and people gonna find loopholes. No?

Would be pretty cool, to use your GitHub as a part of your memorial. Open source the will, with the names anonymized of course.

There's a part of me that would just love for my next of kin to read some of my code to see how I thought when I was alive. Feels like a very interesting tech TV miniseries, an Open Project that's passed down from generation to generation.

I would hope that they would keep accounts open. I have written several books that rely on GitHub for the examples. I am almost 70, and if I should disappear unexpectedly, I would hope that readers would have the code available until the material is no longer relevant.

I can't speak to the legal options available and I am not a lawyer, but if there is a beneficiary you can put in your will/trust that you trust with access to Github, then you could put a couple copies of a keypass database, or even a text file on a couple of USB sticks and put them in a safety deposit box for someone to inherit. I would assume they have power of attorney. Perhaps you can make an agreement with them to periodically push a commit to a README.md file. This assumes you are not making payments to Github. If so, the executor / beneficiary in your will or trust would have to take over making payments. If that is not an option, then maybe the account will stay active if you make them contributors to your project. I don't know if Github will keep a project live if at least one contributor allowed read-write access is active. Everything I am suggesting may violate some AUP/ToS, but I have never been one to let such things get in the way when no other options make themselves available.

On a side note, something I do with VPS providers is to prepay them extra so that I have credit in my account. Even if I stop paying them, my services will stay live for years, if for no other reason than to have cron jobs email people I know so that my morbid bad humor can live on for some time. Perhaps you could do something similar so that your book continues to make money for your estate? Maybe another option would be to find a site that specializes in hosting static content specifically for this purpose and put in redirects on github to that site.

Now we need Steam and other companies that "sell" you digital products to do the same.

I don't want my Steam library to go to waste when I die.

It’s possible and likely that next of kin has no idea what to do with a GitHub account.

Would be better to add a deceased.txt or a text box in settings somewhere.

Can I only have one successor?

If you have multiple heirs, the oldest one is supposed to be entitled as this is an unsplittable item.

Sure, but what if I put someone as an successor and they pass away before me? Wouldn't it be useful to have at least a (prioritised) list? If the first one becomes unavailable (deletes GitHub account, passes away etc) then I could at least be notified by GitHub that I need to update that list.

If you have gavelkind succession then each of your repos goes to a different heir.

Yes. Just sign up your SO as the designated survivor :)

Github suspended my account after 12 years and ignores me in every possible way to try to fix it, even though I know various company that work at the company. Github=Microsoft, Github died the day MSFT bought it.

Reply on news.ycombinator.com