Nice article! We've been doing similar problem solving for years, works fine. 2 things:

- instead of using SSMS to find xact_seqno vlues, we get them by querying distribution.dbo.MSrepl_errors

- rather than editing the replication procedures to ignore errrors (we used to do this) we now adjust the subscription's profille to "continue on data consistency errors" which does just about same thing. When we get past errors, we change back.