If Email had been designed from day one so that you could not send emails with forged addresses, a lot of this stuff would never happen.
A little bit of history
E-mail as we know it (name @ destination) and the network-of-networks idea that became the internet date back to the early seventies. Simple Mail Transfer Protocol (SMTP) and the sendmail program (still very widely used) date back to 1981. ARPANET switched to TCP/IP, the protocol that the internet still uses, in 1982. Back then the idea of spam, e-mail nasties, DoS attacks and forged addresses were inconceiveable - the infant internet was a tightly-knit collection of (mainly American) official networks. Why should they have thought of designing a secure forgery-resistant e-mail protocol? Or SYN-attack resistant IP? Every extra byte cost a lot of computing and network power. E-mail and the internet itself was designed to be as simple as they could get away with.
The problem is that there is a huge amount of inertia in the system - so much so that 20 years later we're still using essentially the same software and protocols, and they're no longer good enough. However, fundamental changes are very hard to make - suddenly networks become isolated from each other due to different protocols and we are temporarily back in the 70s. People have been trying to get IPv6 adopted for years,
and that's a fairly trivial change designed mainly to open up a larger number of IP addresses - the current IPv4 system doesn't really have enough to go around. Developing and rolling out a globally-secure digitally-signed e-mail protocol is a much harder problem to solve.
Solutions? There isn't an easy one. Best bet may be to build a whole new system from the ground up and roll it out in some way. Backwards compatibility is probably the wrong idea, but it makes the migration of millions of non-technical users to "Internet v2.0" a huge problem. Make it backwards compatible and you leave the prospect of all the old holes. It's a problem for someone smarter than me