When email was designed, everybody played nice, and trusted one another. So for a long time, MTAs were open relays out of the box:

Anyone could mail through an open relay to anyone else.

Then, when spam became a problem, people started closing their open relays, so only legitimate mail was accepted.

People started trying to fight spam by blocking certain client IPs which were known spammers. Blacklists became popular and people started using them to identify open relays and dialups as well.

But IP blacklists are an imperfect solution: they result in too many false positives and false negatives. And they're embroiled in lawsuits: spammers are fighting back. People are trying all kinds of desperate moves like renaming them "blocklists". Blacklists aren't working.

And that still doesn't address the forgery issue. SMTP is anonymous on the sender end. Anyone can inject a message to a valid recipient, and forge the sender address. This is called Joe-Jobbing and it happens everyday.

SPF To The Rescue!

SPF takes a different approach: it's the responsibility of each domain owner to whitelist their own mail senders. This prevents forgery, and, because most spam is forged, also blocks spam.

Spam that isn't forged can be easily blacklisted by domain rather than by IP. After widespread adoption, SPF can become even stricter:

This would not have worked in the early days of the Internet, but it is the right solution for today's world.