Reading an article on replication for sql server 2k, it mentioned about
sp_repladdcolum n/sp_repldropcolu mn stored procs, implying that when a
new column is added or dropped from source table, one needs to execute
some sp to ensure that related replication meta data would be in sync.
I'm just thinking loud here, essentially there are two ways to do it
(by design):
* manually making replication meta data match with source table meta
data (as now with 2k server). Big disadvantage: one can easily forget
about this step.
* automatically making replication meta data match with source table
meta data whenever source meta data changes. Some assumptions and cost
for this option:
a) the need to add or drop a column for the source table would also
exist for the replicated data;
b) the mechanism to perform replication meta data update when ddl
occures at source table (price)
So, maybe there's a time one wants to do it manually and other time
doing it automatically would be more desirable. Easily solved if
server/database setup configuration has an option to support that.