Here's a ponder:

What if Gmail's spam filters - which are indeed smart - use a collective behavior analysis to decide whether an email is spam.

Say one filter is whether the email was sent from a server or domain name known to send spam. This seems likely. And how does Google know the server usually sends spam? What if they track how many times the emails are deleted versus how many times emails are kept. Maybe they're tracking time spent on each email, too.

The problem arises when a new server or domain starts sending emails. I'm not talking of emails sent to actual users. I'm talking about emails sent to developers during the development process.

Emails sent to test the contact form, emails sent to test email notifications, emails with debug data... You get the idea.

Most developers simply delete these emails. And rightly so.

What if that makes Google think the site is sending out a large number of unwanted emails - an consequently black-labels the newborn website as a spam server?

And Yahoo! can go suck an acorn with their spam filter.

Meh.