I'm confused then because per this article it seems like it would work - http://msdn.microsoft.com/en-us/library/ms143699.aspx . It gives three rules for the replication topology between versions:
• A Distributor can be any version as long as it is greater than or equal to the Publisher version (in many cases the Distributor is the same instance as the Publisher).
• A Publisher can be any version as long as it less than or equal to the Distributor version.
• Subscriber version depends on the type of publication:
o A Subscriber to a transactional publication can be any version within two versions of the Publisher version. For example: a SQL Server 2005 Publisher can have Subscribers running SQL Server 2005, SQL Server 2008 (including SQL Server 2008 R2), or SQL Server 2012; and a SQL Server 2012 Publisher can have Subscribers running SQL Server 2005, SQL Server 2008 (including SQL Server 2008 R2), or SQL Server 2012.
o A Subscriber to a merge publication can be any version less than or equal to the Publisher version.
And if we are using SQL 2005 with SQL 2012 it appears all three rules would be satisfied. If we ran the Distributor on SQL 2012 and the Publisher on 2005 this seems like it would satisfy the first two rules given the Distributor version is greater than the Publisher version. Also SQL 2005 and SQL 2012 are two versions apart (9.x and 11.x) so I'd think this would satisfy the third requirement. Even in its example it mentions SQL 2005 and SQL 2012.
So am I misreading this article or missing something else?