A Model for Identity in Software | Hacker News Reader

A Model for Identity in Software

christine.website

70 pointstodsacerdoti19 comments

> Mara

> "Real names" is usually a poorly defined concept, however in this case it usually means "whatever is on your government ID", which can be shockingly problematic to transgender or gender-nonbinary people that live in life situations or countries that prevent them from being able to have agency over their government ID.

Not to belittle the problems transgender or nonbinary people face, but forcing real names from government IDs can be problematic even for the average person in many ways. This is not an "it's a problem only for some people" problem. It's also problematic for activists who may face different kinds of threats and real harm. Anonymity or pseudo anonymity can help people express themselves more openly without a constant worry of being shamed and/or banished and/or harmed. On this front, Facebook still stands out as a big abuser where one can easily shutdown someone else's account by reporting it as a fake account. Facebook will then deactivate the account, force the user to provide a government identity document, and in many cases, automatically change the name on the account to the one in the document without asking for consent. Edited to add: Facebook's policies also require that each person create only one account for themselves.

Yeah, I can only really speak to what I've experienced though. If you have any links to descriptions of how "real name" policies harm people I would be more than glad to include them in an upcoming patch to that post.

There is an even more common/simpler problem: what if you have different spellings on different government IDs? Also, what if your name is so long, that it does not fit most gov IDs?

Example: http://entojo.com/14-nomes/ (in Portuguese)

I like this model because I think it really represents what people already do. Articles like this, and ones about privacy (there's some overlap), always focus on LGBT, activists, whistleblowers etc but people in more mainstream groups participate in identity compartmentalization as well. Sometimes they'll open an account for their pet, other times they'll just share less intimate details on more open platforms. I've long thought PGP and derivative keys are a good solution for this.

This isn't a model for identity, this is attempting to justify a rearchitecture of identity systems primarily for solving the scenario of plurality

Many of the things that would ease things up on plural systems would also benefit individuals as well

Author of the post here.

Bingo. This is why I want to introduce the worst case scenario for the existing model and propose a system flexible enough to handle that, this incidentally solves a lot of problems that other people have along the way.

The problem of "how does a computer identify an individual" is clearly a different problem to "how does the individual want to be identified?".

Facebook Work accounts. Teachers running Google Classroom. There are plenty of examples where technical solutions that address the first problem (to make it easier for computers) exacerbate the second (which seeks to make things easier for humans). We should prioritise the technological solutions that help us, not machines.

This makes it sound as though you just skipped over all the stuff about queer people or political dissidents in favor of the thing you found easier to dismiss.

Overall this is an argument for treating human identity humanely, that is, as though the human is the most important part of human identity. That is persuasive to me, particularly in contrast to inhumane models of identity that exist primarily to serve exploitive business models.

I've been thinking about this issue since reading Fall; or, Dodge in Hell by Neal Stephenson. It uses the concept of a Purdah: A pseudo-anonymous online identity. It's related to concepts of trust, fingerprinting, etc.

It inspired me to built a WIP website where you create an identity by linking accounts in various locations, that the site proves you own. The idea is, you build trust / an identity through content you've written on a site - then someone who reads your post on another site knows something about your trustworthiness. I think an especially potent use is in reducing the change a given user (poster, reviewer) etc isn't a bot or marketer.

We'll need to think about this more in the near future as internet communities (and public opinions) are more heavily influenced by bots and other forms of manipulation.

Sort of like RAID - Redundant Array of IDentity

Your site sounds like keybase. While I parked my name there, I have no use for it, because it relies on network adoption. I mean both my friends (for messages) and for other organizations (Sonatype doesn't consider it a gpg key host for signing artifacts). Good luck with your deal.

This was an entertaining survey of whats out there and what problems aren't solved yet. Im sold that we can do so much better than what we have now and Id love to see more explorations!

If you end up making or finding something better, let me know at christine.website/contact. I would be happy to review and give feedback to those kinds of systems.

SSB has a pretty clever name resolution system[1]. The only problem I see with it is that it is a bit too leaky. You may learn someone’s nickname from a friend, but you also may not if your friend doesn’t tell you. SSB makes the gossiping of names a little too default for my taste.

[1]: https://github.com/ssbc/ssb-names/blob/master/README.md

Self sovereign identity, based on decentralised identifiers and verifiable claims (see e.g. https://www.frontiersin.org/articles/10.3389/fbloc.2019.0002...), addresses these (and other) issues systematically.

In my mind that cartoon character smells like a goat.

Mara is a shark, these SFW drawings don't really leave much room for the shark features (fins, body patterns, etc.) to show very prominently.

Reply on news.ycombinator.com