Here’s something I’ve been thinking about that that ties into Joel’s idea that an e-mail service that charged one cent per e-mail would help counteract spammers. I’ve been thinking about the scalability of user-functionality online. From the position of a user the process of sending e-mail scales reasonably well – or at least it basically scales linearly – the incremental cost in effort for each e-mail being dependent only on the elegance of the software you have to automate your e-mail sending process. Adding a financial aspect to sending e-mail doesn’t change the nature of the curve, but it does raise the incremental cost (albeit only financially).
Now there are a variety of reasons why (for example) you’d want to send lots (thousands or tens of thousands) of e-mails, but the main reason will probably be advertising or spam – kinds of e-mail which are generally perceived as an abuse. In fact I would hazard a guess that it becomes increasingly likely that someone is abusing the system the more e-mails they send. (Incidentally – what shape do we think the curve of numbers of e-mails sent vs. numbers of people sending them would be?) So we’re in fact not interested in stopping people sending spam, we’re interested in discouraging people from sending large blocks of e-mail generally…
Essentially what many systems online need – the systems that are prone to abuse that is – are graphs of ‘difficulty of use’ that are exponential – start almost flat and then escalate heavily afterwards. I think this could be true for e-mail (although I suspect it would be impossible to implement outside of a closed system) but also for related processes like posting to message boards or creating online identities… Wouldn’t it be ideal if it was tremendously simple to send up to a hundred e-mails a day, slightly harder to send the next hundred and almost impossible to send ten thousand? We’re not used to thinking in this way, because the dot-com explosion was all about scale and size and making everything work at large-volume and making everything easy for users. But there’s a difference between clarity of function and purpose and ease of use. Maybe there are processes that should be harder generally – and more specifically maybe there are processes that should become harder the more they are used…