> Why can’t we run more things on top of email as a backend?
Security, speed, and lock-in.
Email is hard to secure without using out-of-band mechanisms, and at that point if you're already not adhering to an existing standard you might as well write your own.
Email is fast for what it does, but it's not fast enough to handle the low-latency and interactive environment of something like Twitter. We don't notice this as much because so many people are used to using the same 2-3 email providers (heck, Google alone probably accounts for most personal email), but in a truly federated environment there can be noticeable delays. SMTP explicitly allows for this and provides mechanisms to handle it.
And lastly, building on top of email makes it hard to achieve vendor lock-in, which makes it hard to make money.
Security, speed, and lock-in.
Email is hard to secure without using out-of-band mechanisms, and at that point if you're already not adhering to an existing standard you might as well write your own.
Email is fast for what it does, but it's not fast enough to handle the low-latency and interactive environment of something like Twitter. We don't notice this as much because so many people are used to using the same 2-3 email providers (heck, Google alone probably accounts for most personal email), but in a truly federated environment there can be noticeable delays. SMTP explicitly allows for this and provides mechanisms to handle it.
And lastly, building on top of email makes it hard to achieve vendor lock-in, which makes it hard to make money.