Project-scoped pseudonyms would be a better default for code hosting

I think code-hosting platforms should support project-scoped pseudonyms instead of making one permanent username the center of a person’s technical life. A single profile can connect bug reports, political comments, hobby projects, workplace activity, and long periods of inactivity. Recruiters, data brokers, stalkers, and recommendation systems can all benefit from that graph, even when the user never intended those contexts to meet.

This does not require abandoning accountability. Platforms could preserve verified contribution history, signed commits, moderation records, and an option to carry reputation between identities when the user explicitly chooses it. GitHub already shows how public activity and contribution attribution can exist without exposing a personal email address, so the design space is not “full identity” versus “no trust.”

Pseudonyms can enable abuse, and serious moderation still needs controlled attribution in some cases. But mandatory identity linkage can make newcomers and experimentation pay an unnecessary permanent cost. Has anyone used separate technical identities, or found a better accountability model?

The important distinction is between accountability and universal linkability. Moderators may need a way to investigate abuse, but that does not mean every visitor, recruiter, or ranking system should receive the same identity map by default. I’d want that access narrowly scoped, logged, and contestable.

Project scope sounds clean until someone submits a critical patch across ten dependencies. Do we end up with ten reputations, ten signing keys, and no useful way to tell that the same competent person is behind them? I’m not opposed, but the key-management UX would make or break it.

I’d separate the claims here. Pseudonymisation can make linkage harder while retaining controlled attribution; it does not make linkage impossible. That is still a meaningful privacy improvement, but the threat model should say who can reconnect identities and under what process.

There’s also a social cost to one durable profile: people learn to write for an imagined future audience. Project boundaries could make technical spaces feel more like places to participate and less like permanent auditions.

The abuse concern is bigger than a footnote. Sockpuppets, ban evasion, and reputation laundering become easier when identities are compartmentalized. Maybe the workable compromise is pseudonyms by default, with portable attestations that reveal continuity only when a user opts in.

Separate email addresses and signing keys already get you partway there. The harder problem is behavioral fingerprinting: writing style, timing, interests, and repeated technical quirks. A platform can promise not to link accounts and still leave users exposed elsewhere.

I’d pay attention to the defaults. Optional portability is useful, but many users will click through whatever makes a contribution go live. If privacy is the goal, the platform should make separation easy and make merging an explicit, informed choice—not another buried setting.