In opensource software world it's rather common to hear something like "Does it work just like some-other-application" or "Why can't I use application-X instead of this one, everyone else has application-X".
In a way that's peer pressure. Microsoft is prime example of a company that uses peer pressure for marketing. They managed to acquire a fair share of the market by getting their operating system to OEM markets and by donations to schools. It's interesting that the old saying "There is no such thing as free lunch" applies here too. By donating something they are creating a user base that already knows some application or operating system. By using that user base it is possible to push your products to new markets.
That is actually brilliant marketing.
How is this related to me? Well, some time ago I decided to rebuild my old server. I decided that I would utilize the skills I've learned through my work experience, even if it is a home server. I set up a Xen server that separated my jabber server from my firewall. And since I already had a working IPv6 tunnel and plenty of addresses for my local network I could allow direct connections through IPv6 and forward ports through IPv4. I already knew that XMPP system had already implemented SRV records that allowed me to create clean rule sets how I wanted my servers to be contacted.
But there was a problem. It only occurred to me once I had finished the DNS configuration. It appears that Google Talk doesn't fully support SRV records. It appears that the IPv6 only record throws the servers off and I'm unable to connect to Gtalk users.
So due to peer pressure I was forced to change my preferences on how to connect to my jabber server. Suddenly it became clear to me why it is so hard to introduce new services and technologies. It's not enough if it's innovative or useful, there has to be solid interoperability with the competing products and minimal learning curve. Knowing this, it's not a wonder that the e-mail system is still in place even if it's flawed by design.