• I see, what you mean - when developer A works on a procedure for 4 days and developer B makes a small change (e.g. a new column or fixing a typo) in the same procedure on day two (without telling it), there is a chance, that B's change will be overwritten by A (but at least, we can see this in the SVN history). We have no real solution for this (on the other hand it happend only once or twice in the last few years - we are a small team with only ~10 devs). For non-parallel but near-time changes: each developer has to apply the current deployment scripts (he has just to execute a batch file), before making changes to procedures etc.

    I tested RedGate SourceControl (and products from other companies) for several months before, but they make all to many errors when creating deployment scripts (often not working (mostly because of wrong execution order of altered triggers and created / dropped columns), in the worst case it tried to drop ~ 20 important (and unchanged) tables, just because it decide, that it had to rebuild a verybig table (with a lot of indexes / FKs), just because the formula of a computed column changed a little bit (instead of just dropping / "recreating" this computed column).

    God is real, unless declared integer.